BookPDF Available

Figures

Content may be subject to copyright.
H. J. Temple, S. Anstee, J. Ekstrom, J. D. Pilgrim,
J. Rabenantoandro, J-B. Ramanamanjato,
F. Randriatafika and M. Vincelette.
IUCN and Rio Tinto Technical Series No.2
Forecasting the path towards
a Net Positive Impact on
biodiversity for Rio Tinto QMM
Forecasting the path
towards a Net Positive
Impact on biodiversity
for Rio Tinto QMM
H. J. Temple, S. Anstee, J. Ekstrom, J. D. Pilgrim, J. Rabenantoandro,
J-B. Ramanamanjato, F. Randriatafika and M. Vincelette
IUCN and Rio Tinto Technical Series No.2
The designation of geographical entities in this book, and the presentation of the
material, do not imply the expression of any opinion whatsoever on the part of
IUCN or Rio Tinto concerning the legal status of any country, territory, or area,
or of its authorities, or concerning the delimitation of its frontiers or boundaries.
This publication captures the experience of an innovative and scientifically robust
attempt to assess, forecast, and subsequently account for procedures put in
place by Rio Tinto to mitigate and remediate the input of their operations at the
QMM site.
As the scientific basis and operational procedures to attain a net positive impact
are still in the process of being established, this report is intended to be strictly
technical in nature. It does not represent endorsement of a particular approach
or standard towards attaining “net positive impact”.
IUCN does not take responsibility for the operational delivery of the outcomes
anticipated in the forecast.
Published by: IUCN, Gland, Switzerland and Rio Tinto, London, UK
Copyright: © 2012 International Union for Conservation of Nature and
Natural Resources
Reproduction of this publication for educational or other non-commercial
purposes is authorized without prior written permission from the copyright
holder provided the source is fully acknowledged.
Reproduction of this publication for resale or other commercial purposes is
prohibited without prior written permission of the copyright holder.
Citation: Temple, H.J., Anstee, S., Ekstrom, J., Pilgrim, J.D., Rabenantoandro, J.,
Ramanamanjato, J.-B., Randriatafika, F. & Vincelette, M. (2012). Forecasting the
path towards a Net Positive Impact on biodiversity for Rio Tinto QMM. Gland,
Switzerland: IUCN. x + 78pp.
ISBN: 978-2-8317-1441-7
Cover photo:
© David Rabehevitra
Layout by:
millerdesign.co.uk
Printed by:
SRO-Kundig SA, Switzerland
Available from:
IUCN
(International Union for
Conservation of Nature)
Publications Services
Rue Mauverney 28
1196 Gland Switzerland
Tel +41 22 999 0000
Fax +41 22 999 0020
books@iucn.org
www.iucn.org/publications
The text of this book is
printed on FSC paper.
Contents
Foreword
Acknowledgements
Executive summary
1 Introduction 1
2 Background 5
2.1 National and regional context 5
2.1.1 Madagascar 5
2.1.2 Fort Dauphin region 5
2.2 Physical environment 6
2.3 Biological environment 6
2.3.1 Habitats 6
2.3.2 Species 8
2.4 Rio Tinto QMM project 10
2.5 The Biodiversity Committee 10
2.6 Rio Tinto’s NPI commitment 10
2.7 The mitigation hierarchy 12
3 Methods 15
3.1 Identify and prioritize biodiversity features for inclusion in
NPI accounting 15
3.2 Decide which metrics to use 16
3.3 Select counterfactual scenario(s) against which to
measure losses and gains 17
3.4 Quantify habitat losses for the periods 2004–2015
and 2004–2065 20
3.4.1 Assessing forest condition 21
3.5 Quantify habitat gains for the periods 2004–2015
and 2004–2065 22
3.5.1 Restoration 22
3.5.2 Quality gains in the Avoidance Zones 25
3.5.3 Averted deforestation gains 25
3.5.4 Credit claims 27
3.6 Calculating species losses and gains for the periods
2004–2015 and 2004–2065 27
4 Results 29
4.1 Is Rio Tinto QMM on track to achieve a Net Positive
Impact on biodiversity for the period 2004–2065? 29
4.1.1 Summary 29
4.1.2 Impacts on habitats—Quality Hectares 30
4.1.3 Impacts on species—Units of Global Distribution 35
4.2 Is Rio Tinto QMM on track to achieve a Net Positive
Impact on biodiversity for the period 2004–2015? 40
4.2.1 Impacts on habitats—Quality Hectares 40
4.2.2 Impacts on species—Units of Global Distribution 40
4.3 Is Rio Tinto QMM on track to achieve a Net Positive
Impact on biodiversity throughout the lifecycle of the mine? 41
5 Conclusions and next steps 43
5.1 Summary 43
5.2 Achieving NPI—what does it mean for Rio Tinto QMM? 44
5.3 Rio Tinto QMM’s biodiversity offsets 45
5.4 Monitoring forest loss 47
5.5 The broader context—drivers of biodiversity loss in
southeastern Madagascar 49
5.6 NPI analysis updates and adaptive management 49
6 Lessons learned and directions for future research 50
6.1 Potential impacts of climate change 50
6.2 Defining a standard set of species included in NPI accounting 50
6.2.1 Undescribed species 51
6.3 Counterfactual scenarios, baselines, and calculating
biodiversity gains from averted loss 51
6.4 Calculating Units of Global Distribution—methods for
measuring species distribution area 51
7 Bibliography 53
Appendix 1 61
Appendix 2 62
Appendix 3 64
Appendix 4 66
Appendix 5 72
List of tables
1 Predicted net impact of Rio Tinto QMM 2004–2065 30
2 Losses and gains in QH predicted for 2004–2065 32
3 Animal species with a net negative impact at 2065 37
4 Like-for-not-like gains in site-endemic plant species 38
5 Like-for-not-like gains in site-endemic, Endangered and
Critically Endangered animal species 39
List of figures
1 Map of Rio Tinto QMM project area 3
2 The mitigation hierarchy 13
3 Predicted changes in littoral forest cover 18
4 The three counterfactual scenarios considered 19
5 Littoral forest extent and condition at Rio Tinto QMM in 2005 24
6 Losses and gains of Quality Hectares of forest 2004–2065 33
7 NPI forecast 34
8 Cumulative gains in littoral forest over time 42
9 Current biodiversity offset sites 48
FOREWORD
This publication aims to provide the data, theory, and predictions for the potential
long-term outcome of a biodiversity conservation programme at a mining site.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto
QMM is a joint initiative between IUCN and Rio Tinto, a global mining Group.
Rio Tinto and IUCN signed a three year collaborative agreement in 2010, with the
overall objective to build a business focused collaboration that enables Rio Tinto
to improve its conservation outcomes, strengthen IUCN and Rio Tinto capacities
for market-based approaches to conservation, and contribute to industry-wide
improvements in the mining and associated sectors. By working together, both
organizations aim to better understand each other’s issues and priorities, draw
on each other’s experience and expertise, and develop programmes and actions
that provide on-the-ground conservation value and contribute to improved
performance – for IUCN, Rio Tinto, and the mining sector more broadly.
This report contributes to the ongoing discussion around applying a mitigation
hierarchy to managing biodiversity risks and challenges and discussion around
biodiversity offsets and biodiversity metrics. IUCN and Rio Tinto believe an
important input to the discussion is rigorous on-the-ground testing of theories
and methodologies, which is robustly and transparently documented.
IUCN is the world’s oldest and largest global environmental organisation, with
more than 1200 government and NGO members and almost 11 000 volunteer
experts in some 160 countries. IUCN’s work is supported by more than 1000
staff in 45 offices and hundreds of partners in public, NGO and private sectors
around the world. IUCN works on biodiversity, climate change, energy, human
livelihoods and greening the world economy by supporting scientific research,
managing field projects all over the world, and bringing governments, NGOs,
the UN and companies together to develop policy, laws and best practice.
Rio Tinto is a leading global mining group, involved in every stage of the mining
business. Its interests are diverse both in geography and product and it works
in some of the world’s most difficult terrains and climates. Most of Rio Tinto’s
assets are in Australia and North America, but it also operates in Europe, South
America, Asia and Africa. Businesses include open pit and underground mines,
mills, refineries and smelters as well as a number of research and service facilities.
Rio Tinto comprises five principal product groups - Aluminium, Copper, Diamonds
& Minerals, Energy, and Iron Ore – plus two support groups: Technology &
Innovation, and Exploration.
Clare Verberne, Dennis Hosack, and Stuart Anstee
ACKNOWLEDGEMENTS
This report has been written by The Biodiversity Consultancy Ltd., Cambridge,
UK, with contributions from the QMM Biodiversity Team and Rio Tinto HSE.
The first draft of this report was prepared by Helen Temple (The Biodiversity
Consultancy), with input from the Rio Tinto QMM biodiversity team (Manon
Vincelette, Johny Rabenantoandro, Faly Randriatafika, Jean-Baptiste
Ramanamanjato, Dominique Andriambahiny) and Jonathan Ekstrom (The
Biodiversity Consultancy) gathered in workshops held in Fort Dauphin, Madagascar
in January and February 2010. At this stage, the document was intended as a
discussion paper to inform debate with the Rio Tinto QMM Biodiversity Committee
regarding what ‘Net Positive Impact’ meant in practice for Rio Tinto QMM. The
methodology and draft text was reviewed by the Rio Tinto QMM Biodiversity
Committee at a workshop held on 3-6 May 2010 in Fort Dauphin (a full day
was dedicated to this task). The following committee members were present
and contributed to the discussion: Porter P. Lowry (Missouri Botanical Gardens);
Jörg Ganzhorn (Hamburg University); Alison Jolly (Sussex University); Rob Brett
(Fauna and Flora International); Paul Smith (Royal Botanic Gardens, Kew); and
Lisa Gaylord (Wildlife Conservation Society). A number of significant changes
and improvements were made to the methodology and draft report based on
the Committee’s requirements. The document was then internally reviewed by
Rio Tinto Health, Safety, Environment and Communities (Stuart Anstee), Rio
Tinto QMM (Manon Vincelette, Johny Rabenantoandro) and The Biodiversity
Consultancy (John Pilgrim, Jonathan Ekstrom). At this stage it was decided that
the document would be of value to a number of external audiences and should
be formally published. Consequently a further round of review and improvement
was initiated in December 2010, involving the Rio Tinto QMM Biodiversity
Committee and a number of additional experts. These experts were selected by
IUCN rather than Rio Tinto QMM, and were: Thomas Brooks (NatureServe), Sue
Mainka (IUCN), Dennis Hosack (IUCN), Monica Barcellos-Harris (UNEP-WCMC),
Leon Bennun (BirdLife International), Conrad Savy (Conservation International)
and James Watson (Wildlife Conservation Society). In addition, the document
was reviewed at the first meeting of the IUCN-Rio Tinto Working Group on NPI
Verification, held on 17-18 March 2011 in London, attended by Stuart Anstee
(Rio Tinto), Rachel Asante-Owusu (IUCN), Dennis Hosack (IUCN), Sally Madden
(Rio Tinto), Thomas Brooks (NatureServe), Manon Vincelette (Rio Tinto), Conrad
Savy (Conservation International), Simon Wake (Rio Tinto), Monica Harris (UNEP-
WCMC), James Watson (Wildlife Conservation Society) and Rainer Schneeweiss
(Rio Tinto). Finally, the document was also graciously reviewed by Richard Jenkins
(IUCN) and Paul Racey (University of Exeter).
The authors should like to thank all of them for their contributions to this report.
EXECUTIVE SUMMARY
Rio Tinto is committed to achieving a Net Positive Impact (NPI) on biodiversity,
a strategy launched at the 2004 IUCN World Conservation Congress. The Rio
Tinto ilmenite mine in southeastern Madagascar, run by QIT Madagascar Minerals
(Rio Tinto QMM), has been chosen as a pilot site to test the tools designed to
achieve and quantify NPI on biodiversity. The most important direct negative
biodiversity impact resulting from the Rio Tinto QMM operation is the loss of
littoral forest habitat at the Mandena, Petriky and Sainte Luce (hereafter Ste
Luce) mining sites. Approximately 1,665 ha (3.5% of Madagascar’s remaining
47,900 ha of littoral forest) are expected to be lost over the next 40 years as a
result of mining and associated activities. Rio Tinto QMM operates a dredge mine
rather than a conventional open cast mine so habitat loss will be incremental
over decades as the mine moves slowly through the landscape; the total direct
footprint is anticipated to be c.8,000 ha over the mine’s lifetime; however the
mine itself occupies c.50 ha at any one time (covering c.100 ha per year: Rio Tinto
QMM, 2001). Rio Tinto QMM began its mining activities in 2009. Littoral forest
has been identified as a national conservation priority (Ganzhorn et al., 2001)
owing to its limited extent and its high concentrations of nationally and locally
endemic plant species (Du Puy and Moat, 1998; Dumetz, 1999), diverse tree flora
(Dumetz, 1999), and high diversity of fauna (Ganzhorn, 1998; Ramanamanjato
et al., 2002; Watson et al., 2005). Littoral forests on the mining concession
harbour many restricted-range species and species classified as Threatened on
the IUCN Red List, including 42 plants and at least 14 invertebrate species that
are found nowhere else in the world.
In the present analysis, biodiversity losses and gains were measured and forecast
for the period 2004–2065 (i.e. from the date of the NPI commitment to the
anticipated date of mine closure), in order to determine whether the current and
proposed mitigation activities of Rio Tinto QMM are sufficient to achieve NPI
by closure. NPI was defined (in consultation with Rio Tinto QMM’s Biodiversity
Committee) as Net Positive Impact on littoral forest (measured in Quality Hectares,
QH) and Net Positive Impact on High Priority species (measured in Units of
Global Distribution, UD).
Four main types of conservation actions are being implemented by Rio Tinto
QMM to mitigate project impacts on key habitats and species. These are:
• Avoidance—at Mandena, Petriky and Ste Luce. Avoidance Zones (AZ) have
been established. They represent a cost to Rio Tinto QMM of c.8% of
foregone ilmenite, as well as the management cost of maintaining these
areas, and protect 27% of the best quality remaining forest cover on the
deposit. Collectively, these cover an area of 624 ha.
The Rio Tinto ilmenite
mine in southeastern
Madagascar, run by
QIT Madagascar
Minerals, has been
chosen as a pilot site to
test the tools designed
to achieve and
quantify a NPI on
biodiversity.
• Minimization—reduction of the likelihood or magnitude of biodiversity
impacts from mining activities that cannot be avoided. At Rio Tinto QMM
this includes a diverse range of activities such as minimizing disturbance
and roadkill from mining-related traffic by educating drivers and enforcing
strict speed limits.
• Rehabilitation and restoration—re-establishment of littoral forest on areas
that have been completely cleared, by replacing topsoil (stored during the
mining process) and planting with appropriate native species propagated
in Rio Tinto QMM’s nursery. There are plans for the restoration of c.225
ha at each of the three sites (Mandena, Petriky and Ste Luce), amounting
to c.675 ha in total. Restoration zones will be located adjacent to the
Avoidance Zones, to provide a buffer, improve connectivity and facilitate
natural regeneration and re-colonization.
• Biodiversity offsets—Rio Tinto QMM is investing in biodiversity offsets
at several forest sites in the region, with the aim of reducing the high
background rate of deforestation. These offset sites cover c.6,000 ha
of forest.
In addition, Rio Tinto QMM is carrying out a number of additional conservation
actions (e.g. environmental education, capacity-building, livelihoods alternatives,
etc.) with the aim of making a positive contribution to sustainable development
in the region and reducing human pressure on biodiversity.
Net impact on littoral forest is forecast to be +350 QH in 2065,
1
representing an
increase in forest extent and quality of 13% in comparison to 2004 (measured in
QH). Net impact on littoral forest is forecast to be +48 QH in 2015, representing
an increase of 2% in comparison to 2004 levels. Net impact on the forest as
a whole (including the humid forests of the Bemangidy
2
offset) is forecast to
be +1,251 QH.
Loss of littoral forest caused by direct impacts of mining is predicted to be -428
QH. Total littoral forest gain is predicted to be +778 QH. Consequently the ratio
of gain to loss is approximately 2:1. Considering all forest types, loss remains
constant at -428 QH; gain in all forest types (including Bemangidy humid forest)
is +1,679 QH. In this case the ratio of gain to loss is approximately 4:1.
Of the 90 High Priority terrestrial species (54 plants, 26 invertebrates, 10
vertebrates) individually, 83 (92%) are forecast to show a Net Positive Impact
at 2065. For 59 of these, area-based calculations predict that NPI will be reached;
for a further 24 plant species, area-based calculations predict a moderate
negative impact (-1.3% to -17.9%) requiring Rio Tinto QMM to reach NPI by
2065 through enrichment of the avoidance and restoration zones (e.g. returning
species from their current depleted levels towards estimated natural densities in
1 Based on a background deforestation rate of the Madagascar national average for c.1990 to c.
2000 of 0.9% p er year.
2 Bemangidy is part of the Tsitongambarika forest – it lies within the area of the forest that is
sometimes called ‘TGK III.
optimal conditions
3
). Seven animal species show residual negative impactsfour
vertebrates (including the Critically Endangered gecko Phelsuma antanosy) show
residual negative impacts of up to 5.1%, and three invertebrates show residual
negative impacts of up to 17.9%. It is not known whether enrichment would
be feasible or desirable for these species,4 research is underway to investigate
options.
Only 32 of the 90 High Priority species are likely to be at NPI by 2015, and some
species show losses of up to 39% of their global distribution at this point (two
millipedes and one plant, all endemic to Mandena, show population reductions
of this magnitude). This is because there will already have been major impacts
by this date (particularly at Mandena), but rehabilitation/restoration zones will
not as yet have reached sufficient maturity to generate gains that can count
towards NPI.
Importantly, this analysis is primarily a forecast based on the best available data.
Accounting of biodiversity losses and gains will be required periodically, using
the same principles and methods, and incorporating better information as it
becomes available.
Overall, this analysis shows that Rio Tinto QMM could be on track to achieve a Net
Positive Impact on biodiversity by the date of closure of the mine, provided that:
1. The assumptions upon which the analysis is based (inter alia background
deforestation rate, levels of habitat degradation, rate at which habitat
quality can be restored) are either accurate or precautionary.
2. The conservation measures detailed here are successfully implemented.
3. Further research is conducted into conservation options for the four
vertebrate and three invertebrate species predicted to have a net negative
impact in 2065, and the best of these options is implemented as a high
priority.
4. Sustained investment in conservation action is assured.
5. Rigorous monitoring and independent verification are implemented to
ensure that real biodiversity gains are achieved.
3 At present, littoral forest at Mandena, Petriky and Ste Luce is somewhat degraded as a result of
various human activities (not only mining)—consequently, for some species (e.g. plant species
highly sensitive to disturbance or degradation) it is desirable to ultimately aim to restore them to
greater densities than are currently observed.
4 Note that Phelsuma antanosy is found at Ambatotsirongorongo, which has been protected
through a joint QMM-Wildlife Conservation Societ y initiative. The ‘gains’ from this site are not
included in the summary figures presented here.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 1
1 INTRODUCTION
Rio Tinto is committed to achieving a Net Positive Impact (NPI) on biodiversity at
sites where it operates, a strategy launched at the 2004 IUCN World Conservation
Congress and reinforced at the 2008 Congress (Rio Tinto, 2008a). The Rio
Tinto ilmenite mine in the Fort Dauphin region of southeastern Madagascar
run by QIT Madagascar Minerals (Rio Tinto QMM) has been chosen as one of
Rio Tinto’s pilot sites to test the tools designed to achieve and quantify NPI on
biodiversity. It consists of three sites to be mined sequentially (Mandena, Ste
Luce and Petriky), a new deepwater port, and ancillary infrastructures such as
roads, quarry, housing and industrial areas (Figure 1). Mining at the first of
these three sites, Mandena, began in 2009. Rio Tinto QMM has made a formal
commitment in its Biodiversity Action Plan to achieve NPI on biodiversity.
The most important direct negative biodiversity impact resulting from Rio Tinto
QMM’s activities is the loss of littoral forest habitat at Mandena, Petriky and Ste
Luce. Littoral forest is a rare and threatened habitat within Madagascarc.90%
of this habitat type has already been lost as a result of human activities (Consiglio
et al., 2006). Approximately 1,665 ha (3.5% of Madagascar’s remaining 47,900
ha of littoral forest) is expected to be lost to dredging, which entails not only
clearance of vegetation but also removal of soil and its constituent seed bank.
Littoral forests on the mining concession harbour many restricted-range species
and species classified as Threatened by the IUCN Red List, including 42 plants
and at least 14 invertebrate species that are found nowhere else in the world.
The project will have substantial residual impacts on a number of these species.
Four main types of conservation actions are being used by Rio Tinto QMM to
mitigate project impacts on key habitats and species. These are:
• Avoidance Zones (AZ)—at Mandena, Petriky and Ste Luce, Avoidance
Zones have been established on the ilmenite deposits to protect those
blocks of littoral forest that are in the best condition. These AZ have been
officially incorporated into Madagascar’s national Protected Areas network.
They represent a cost to Rio Tinto QMM of c.8% of foregone ilmenite and
protect 27% of forest cover on the deposit. Collectively, they cover a total
area of 624 ha.
• Minimization—reduction of the likelihood or magnitude of biodiversity
impacts from mining activities that cannot be avoided. At Rio Tinto QMM
this includes a diverse range of activities such as minimizing disturbance
and roadkill from mining-related traffic through road-safety awareness
campaigns and by enforcing strict speed limits.
Rio Tinto is committed
to achieving NPI on
biodiversity at sites
where it operates, a
strategy launched at
the 2004 IUCN World
Conservation Congress
and reinforced at the
2008 Congress.
2 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
• Rehabilitation and restoration—attempts to re-create littoral forest
on areas that have been completely cleared, by replacing topsoil (stored
during the mining process) and planting with appropriate native species
propagated in Rio Tinto QMM’s nursery. Rehabilitation/restoration zones
will be located adjacent to the AZ to provide a buffer, improve connectivity
and facilitate natural regeneration and re-colonization.
• Biodiversity offsets—Rio Tinto QMM is investing in biodiversity offsets
at several forest sites in the region, with the aim of reducing the high
background rate of deforestation. These offset sites cover c.6,000 ha
of forest.
In addition, Rio Tinto QMM is implementing a number of additional conservation
actions (environmental education, capacity-building, livelihoods alternatives,
etc.) intended to make a positive contribution to biodiversity conservation and
sustainable development in the region. These actions will be essential to reduce
human pressure on remaining forest and to allow restoration, biodiversity offsets
and avoidance to deliver conservation gains.
This report quantifies the current and projected direct impacts on key habitats
and species caused by mining and associated activities, and the potential
conservation gains that may be achieved through habitat restoration and averted
deforestation at Avoidance and Offset Zones. It is intended to shed light on the
following questions:
1. Is Rio Tinto QMM on track to achieve a Net Positive Impact on biodiversity
for the period 2004-2065?
2. Is Rio Tinto QMM on track to achieve a Net Positive Impact on biodiversity
by 2015?
The 2004 to 2065 time frame has been chosen as it corresponds to the period
starting from the launch of the biodiversity NPI policy in 2004 to the current
projected date of mine closure and relinquishment of the tenement in 2065.
The 2015 date represents an interim milestone that has been set as part of Rio
Tinto’s performance target-setting process.
It should be noted that the complexity associated with setting and monitoring
biodiversity metrics means that value judgements are often required in the initial
establishment of relevant targets for achieving NPI. In the case of the Rio Tinto
QMM project the validity of these value judgements was tested by an external
committee of biodiversity and conservation specialists.
5
Quantitative analyses
such as those presented here can provide useful background information and
5 A Biodiversity Advisory Committee was formed in 2001 to review the Rio Tinto QMM biodiversity
strategy and conservation measures on the ground. It consis ts of biodiversity experts with
extensive experience and global renown for their research in Madagascar. This commit tee currently
comprises : Dr. Porter P. Lowry (Missouri Botanical Gardens); Prof. Joerg Ganzhorn, (Hamburg
Universit y); Prof. Alison Jolly (Sussex Univer sity); Dr. Rob Brett ( Fauna & Flora International); Dr.
Paul Smith (Royal Botanic Gardens, Kew); and Lisa Gaylord (Wildlife Conservation Society).
Rio Tinto QMM is
implementing a
number of additional
conservation actions
intended to make a
positive contribution to
biodiversity conservation
and sustainable
development in
the region.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 3
insights, but they should not replace consultation with appropriate biodiversity
stakeholders and experts.
Finally, it should be noted that this report focuses on technical biodiversity issues
because its remit is to answer specific technical questions. It only very briefly
touches on local communities and broader sustainable development issues. This
should not be taken as implying that “biodiversity issues are more important
than social issues for Rio Tinto QMM”, rather it reflects the specific aims of the
report. It is essential that mining and implementation of mitigation and offset
measures are done in a way that takes into account the needs and rights of
local communities and does not leave them worse off than before. Faced with
development of a mine and development of biodiversity offsets, there is a real
risk that local communities may face a ‘double whammy’ of negative impacts
from both initiatives (e.g. if a community is dependent upon forest resources,
and its access to forest is reduced through mining-caused deforestation and
the implementation of a ‘fortress-style’ protected area)(BBOP, 2009). Rio Tinto
QMM’s expressed aim is to implement mitigation and offsets in a way that
benefits both biodiversity and local communities (Rio Tinto QMM, 2010).
4 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Figure 1.
Map of Rio Tinto QMM project area, showing the three deposits
(Mandena, Petriky, Ste Luce) and remaining littoral forest cover on
the deposits in 1998.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 5
2 BACKGROUND
This section provides an overview of the environmental context for Rio Tinto
QMM’s operation in the Fort Dauphin region of southeastern Madagascar. It
also summarises Rio Tinto’s Net Positive Impact commitment and outlines the
mitigation hierarchy, a conceptual tool used by Rio Tinto and others to think about
impacts and mitigation measures. A large number of technical reports, scientific
papers, a Social and Environmental Impact Assessment6 (Rio Tinto QMM, 2001),
a Biodiversity Action Plan (Rio Tinto QMM, 2010) and a 400-page monograph
(Ganzhorn et al., 2007. Biodiversity, ecology and conservation of littoral ecosystems
in southeastern Madagascar) have been compiled by Rio Tinto QMM and the
many scientists who have worked at the mine site and its environs. This section
does not replicate this detailed information but rather gives a concise summary
and directs the reader to sources where more detailed information can be found.
2.1 National and regional context
2.1.1 Madagascar
Madagascar is the fourth largest island in the world, covering 587,000 km2, about
the size of Texas or France. The country is a global biodiversity hotspot with a
rich and unique biodiversity that is subject to high levels of threat (Mittermeier
et al., 2004). Madagascar is home to 12–14,000 vascular plant species, 90%
of which are endemic and found only at a few sites (Mittermeier et al., 2004).
The island has 340 native species of reptile, including more than half the world’s
chameleon species (Raxworthy, 2003). There is almost 100% endemism among
Madagascar’s 222 amphibian species. The country is characterized by high rates
of poverty, large rural populations, subsistence agriculture, and low levels of
industry (Vincelette et al., 2007a). It ranks amongst the poorest of the world’s
countries with per capita GDP estimated at US$221 in 2003 (Vincelette et al.,
2007a) and US$392 in 2010 (IMF, 2011).
2.1.2 Fort Dauphin region
The mine is situated near the town of Fort Dauphin in the Anosy region of
southeastern Madagascar. This is one of the most ecologically diverse regions
of Madagascar (Goodman and Ramanamanjato, 2007), but also one of the
poorest and most isolated. Eighty-two per cent of Anosy inhabitants live below
the poverty line (US1$/day) and the regional population is expected to double by
2020 (Vincelette et al., 2007a). The low-level commercial economy is supported
by just three main productsrice, sisal and lobsterand there is growing pressure
on the environment from unsustainable subsistence use of natural resources
(Vincelette et al., 2007b)7.
6 This SEIA covers the Mandena site. SEIAs for the other sites ( Petriky and Ste Luce) will be carried
out in the future (and in advance of any mining activity at those sites).
7 Further information on the national and regional context can be found in the Social and
Environmental Impact Assessment (SEIA) ( Rio Tinto QMM, 2001) and in V incelette et al. (2 007a ).
Madagascar is a global
biodiversity hotspot
with a rich and unique
biodiversity that is
subject to high levels
of threat. It is home to
12–14,000 vascular
plant species, 90%
of which are endemic
and found only
at a few sites.
6 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
2.2 Physical environment
The Fort Dauphin region is dominated by the Vohimena Mountains and a rolling
coastal plain that extends down to the Indian Ocean. This plain is composed
mainly of littoral sands (often mineralized) that form a series of low dunes
terminating at the shoreline in a series of coastal lagoons. The regional climate
is warm and humid, with occasional cyclones. Average monthly temperatures
range from 26.9°C in January to 20.3°C in July. January is typically the wettest
month and September the driest. Annual precipitation shows a steep gradient
from Petriky in the south (the driest of the three sites) to Ste Luce in the north.
Mandena, located between Petriky and Ste Luce, has a mean annual rainfall
of about 1,600 mm.8
2.3 Biological environment
2.3.1 Habitats
Southeastern Madagascar contains a significant diversity of natural forest habitats
within a complex topographic relief, with few parallels in any other region of the
island. Forest types include coastal littoral forests on sandy substrates, humid
forest habitats ranging from lowland to montane formations, dry forest and
spiny bush. A diverse biota is associated with these habitats, many species and
sub-species of which are locally endemic (Rio Tinto QMM, 2001; Goodman and
Ramanamanjato, 2007). The north-south aligned Anosyenne Mountains act as
a major barrier for weather systems reaching the island from the east. There
are very abrupt ecotones9 on the western flank of this chain on account of this
rain shadow effect, including one of the most extreme known in the world on
the western flank of the Anosyenne Mountain chain between parcels I and II
of Andohahela National Park (Nussbaum et al., 1999). Here, evergreen humid
forest characteristic of the east coast mountain ranges merges into the spiny
sub-desert scrub characteristic of southwestern Madagascar over a remarkably
short distance of about 5 km. For those unfamiliar with Madagascan vegetation
types, this is equivalent to a change from ‘rainforest’ to ‘scrub or maquis-type’
habitat, a change of structural and compositional magnitude with few parallels
globally (Goodman and Ramanamanjato, 2007).
2.3.1.1 Littoral forest
A particularly important terrestrial habitat type found in the mining zone is
littoral forest. Madagascan littoral forests are a sub-type of humid and sub-
humid evergreen forest occurring on sandy substrates (Rabevohitra et al., 1998;
de Gouvenain and Silander, 2000). Littoral forest is notable for its high floristic
diversityalthough it originally occupied less than 1% of Madagascar’s land
surface, 13% of the island’s total native flora has been recorded from this habitat
type (Consiglio et al., 2006). Littoral forest is thought once to have formed a
8 For further information on geology, hydrology and climate in the F ort Dauphin region, see
Vincelette et al. (20 07c).
9 An ecotone is a transitional area between two distinct habitats, where the ranges of the
organisms in each bordering habitat overlap.
Southeastern
Madagascar contains a
significant diversity of
natural forest habitats
within a complex
topographic relief,
with few parallels in
any other region
of the island.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 7
continuous 1,600 km strip along most of Madagascar’s eastern seaboard, however
only c.10% of this remains in the form of small fragments, with only 1.5%
included within the existing Protected Areas network (Consiglio et al., 2006).
Prior to human disturbance, the coastal region including the mining zone
is believed to have been covered in coastal littoral forest (Lowry and Faber-
Langendoen, 1991; Consiglio et al., 2006; but see Virah-Sawmy, 2009). By 1950,
when the first known aerial images of the area were taken, forest cover was
already fragmented and patchy;10 between 1950 and 2005 forest cover in the
Fort Dauphin region further declined by over 50% (Vincelette et al., 2007b).11
At the present time, the mining zone is made up of littoral forest fragments
of varying size and quality, interspersed with highly degraded vegetation, bare
sand, agricultural and inhabited land, and stands of exotics (e.g. Eucalyptus sp.)
and alien invasive tree species (e.g. Melaleuca quinquenervia) (Rio Tinto QMM,
2001; Vincelette et al., 2007a; Rabenantoandro et al., 2007).
In 2005, 3,128 ha of coastal littoral forest remained in the mining zone (Mandena,
Petriky, and Ste Luce; Vincelette et al., 2007b). Since only 47,900 ha of this
habitat remain in the whole of Madagascar (Consiglio et al., 2006), the mining
zone’s forests represent 6.5% of the residual area of this distinctive and highly
floristically diverse habitat type.
2.3.1.2 Comparison of the three sites—Mandena, Petriky and Ste Luce
Dumetz (1999) classified the three southeastern forests (Mandena, Petriky and
Ste Luce) as a unique sub-type of littoral forest on sand. Each of these three
sites has distinct social, physical and ecological characteristics despite their close
geographic proximity to one another (Ingram and Dawson, 2006).
One of the distinctive features of Ste Luce, by comparison with the two other
sites, is that it contains relatively large tracts of littoral forest that remain in fairly
good condition. For example, parcel S912 shows all the characteristics of nearly
intact low elevation dense humid forest, with about 60% cover among trees
that are 12 m or more in height, and a clearly stratified structure (Lowry and
Faber-Langendoen, 1991; Rabenantoandro, 2001; Rabenantoandro et al., 2007).
10 The extent to which this is due to natural fac tors versus anthropo genic factors is debated in the
literature (see e.g. Virah-Sawmy, 2009). The landscape in the Fort Dauphin region may well be
naturally a mosaic habitat (Virah-Sawmy, 2009), but there is also evidence of anthropogenic loss
and degradation dating from before the arrival of QMM (Vincelette et al., 2007b; Virah-Sawmy,
2009) , although this was potentially mainly caused by immigrants from other par ts of Madagascar
rather than by local people and may have been exacerbated by QM M’s exploration-related
activities since the 1990s (Ingram and Dawson, 200 6; Virah-Sawmy, 2009 ).
11 It is possible that, in more recent years, (e.g. 1990s and 200 0s) the rate of loss has been
exacerbate d by the presence of the mining project. However, comparison of imagery from
c.1950, 1972 and 1989 indicates that there was considerable loss of forest cover over this earlier
time period (Figures 2 & 4 in Vincelette et al., 2007; e.g. from c.7,00 0 ha in 1950 to c.4,500 ha
in 19 89) .
12 All parcels of littoral forest at Mandena, Petriky and Ste Luce have been mapped and given
individual numbers.
Between 1950 and
2005 forest cover in
the Fort Dauphin
region further declined
by over 50%.
8 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
At Mandena, the bioclimatic factors are nearly identical to those at Ste Luce,
although precipitation is slightly less. However, remnant parcels of forest are
smaller and in poorer condition, with lower canopy height, smaller trunk
diameters and less stratification. The observed differences in structure found
at Mandena indicate that the forest present there today is a degraded form of
a vegetation type that is shared with Ste Luce. Located just 9 km north of the
town of Fort Dauphin, Mandena has clearly been heavily impacted by humans.
By comparison with Ste Luce, Mandena shows a striking lack of individual trees
belonging to families that are widely used for their wood, such as Ebenaceae,
Sapotaceae and Lauraceae. This is likely a reflection of the previous exploitation
of economically valuable species (Rabenantoandro et al., 2007).
Floristically and faunistically, Petriky can be interpreted as a transition between
dry forest and humid littoral forest (Rabenantoandro et al., 2007). Located
at the extreme southern end of Madagascar’s east coast, Petriky has species
characteristic of humid formations, such as Intsia bijuga (Fabaceae), Homalium
axillare (Flacourtiaceae), Asteropeia multiflora (Asteropeiaceae), and Beilschmiedia
madagascariensis (Lauraceae) but differs from the other two sites through the
presence of taxa typical of dry areas, including Oplonia vincoides (Acanthaceae),
Folotsia madagascariense (Asclepiadaceae), Deinbolia boinesis (Sapindaceae) and
Cordia caffra (Boraginaceae). Similarly, the fauna of Petriky shows clear affinities
with those of dry forest areas in southwestern Madagascar (Ramanamanjato et
al., 2002). One other notable characteristic of the Petriky forest is the lack of
members of the Arecaceae and Pandanaceae families, which are prominent in
the Mandena and Ste Luce forests (Rabenantoandro et al., 2007).
2.3.2 Species
2.3.2.1 Terrestrial species—vascular plants
Of the 614 vascular plant species and varieties recorded from remnant littoral forest
in the mining zone, 83% are endemic to Madagascar, of which 54% are shared
with low- and mid-elevation humid forests, 7% are restricted to southeastern
Madagascar littoral forests,13 6% are restricted to scattered small remnants of
regional littoral forests in the zone between Petriky and Manantenina, and 7%
are found only in the mining zone (Rabenantoandro et al., 2007).14 The number
of plant species strictly endemic to the mining zone currently stands at 42.15
2.3.2.2 Terrestrial species—vertebrates
About 168 species of reptiles and amphibians are found in the Anosy region,
representing around a third of the total herpetofauna of Madagascar (Goodman
and Ramanamanjato, 2007). Ninety-six of these are found in the mining zone.
13 E.g. forests between Mananjary and Fort Dauphin.
14 See Rabenantoandro et al. (2007) for fur ther details and a complete vascular plant species list.
15 The exact number changes over time as a result of research and t axonomic revisions.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 9
Species richness is highest at Ste Luce (69 species), followed by Mandena (63)
and Petriky (45) (Ramanamanjato, 2007).
The Fort Dauphin region exhibits a particularly high bird species richness,
reflecting the exceptional habitat diversity in the Anosy region. Goodman et
al. (1997) listed 189 species occurring in the area south of Manantenina and
east of the Mandrare River, representing 68% of the island’s known avifauna at
that time. Within the mining zone, Watson (2007) recorded 77 bird species in
littoral forest fragments, and Watson et al. (2005) describe these fragments as
holding a unique assemblage of avian species, including both humid and spiny
forest-dependent species, a combination found nowhere else on the island.
However, endemism16 in the region is lowthere is only one regional endemic,
Bluntschli’s Vanga Hypositta perdita, which has not been recorded in the mining
zone (Goodman and Ramanamanjato, 2007; Watson, 2007).
The southeastern portion of the island has a rich small mammal fauna, owing
to the varied habitats in the region and the high mountains of the Anosyenne
and Vohimena Mountains. No endemic species of extant small mammal are
known from the Anosy region. As with other groups of organisms, there are two
principal gradients that show high levels of species turnover within this region: an
east-west gradient from humid forests to dry forests and an elevational gradient,
particularly in parcel I of the Parc National d’Andohahela, from lowland habitats
to sclerophyllous forest in the higher areas. However, compared with other
Malagasy forest types, the littoral forests are depauperate in large mammals.
The Malagasy Ring-tailed Mongoose Galidia elegans and Malagasy Civet Fossa
fossana, familiar inhabitants of many Madagascar ecosystems, are present. In
addition, the Fossa Cryptoprocta ferox, the island’s top predator, was recorded
in Mandena for the first time in 2004. Both micro- and mega-chiropterans have
been recorded in the project area. Several of the mega-chiropteran species roost
at locations near the mining zone, and are believed to be important dispersers
of seeds of the littoral forest. Additionally, the zone has an interesting collection
of primate species. There is at least one restricted-range form of lemur, Eulemur
(fulvus) collaris present in southeast Madagascar and within the mining zone.17
2.3.2.3 Terrestrial species—invertebrates
Invertebrate groups surveyed at the Rio Tinto QMM site to date include
dragonflies (Odonata: Schütte and Razafindraibe, 2007), mantids (Mantodea:
Schütte, 2007), stick insects (Phasmatodea: Schütte, 2007), giant pill-millipedes
(Sphaerotheriida: Wesener and Wägele, 2007; Wesener, 2009), and spirobolid
millipedes (Spirobolida: Wesener et al., 2009).
16 ‘Endemism’ here refers to Anosy regional endemics; se e Goodman and Ramanamanjato (2007)
for further d etails.
17 For more details on the vertebrate species found at QMM’s sites, see Ganzhorn (2007) and
references therein.
The Fort Dauphin
region exhibits a
particularly high
bird species richness,
reflecting the
exceptional habitat
diversity in the
Anosy region.
10 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Based on these surveys, the littoral forests within the mining zone at Mandena,
Petriky and Ste Luce hold a diverse array of invertebrate species, including a
number of species that are endemic or near-endemic to the mining zone.18 The
surveys resulted in the discovery and description of a number of new species,
including the genus Riotintobolus.
2.4 Rio Tinto QMM project
The Rio Tinto QMM project will mine ilmenite ore as a mineral sand, to provide
titanium dioxide to the world market. Titanium is a major ingredient in steels
and other alloys; titanium dioxide is the white pigment found in most paints
and plastics. The Rio Tinto QMM project consists of three sites to be mined
sequentially (Mandena, Ste Luce and Petriky) over a period of c.40–50 years. A
new deepwater port has been constructed at Fort Dauphin. Ancillary infrastructure
includes a dedicated port industrial zone, road networks, housing areas and a
stone quarry. Ilmenite is mined using a dredge situated on artificial ponds that
moves across the ore body as mining progresses; the mining process entails the
removal of all vegetation cover along with the soil. Approximately 100 ha of the
deposit will be mined each year; the mine itself occupies about 50 ha of land
as it progresses slowly through the deposit area. Rehabilitation of the mined
area will be carried out once the dredge has moved on to the next part of the
deposit. The total mine footprint at all three sites collectively is about 8,000 ha.
Ore processing is minimal and takes place on site through physical separation.
Processed ore is transported by truck to the port for export.19
2.5 The Biodiversity Committee
A Biodiversity Advisory Committee was formed in 2001 to review the biodiversity
strategy and conservation measures on the ground. It consists of biodiversity
experts with extensive experience and global renown for their research in
Madagascar. This committee currently comprises: Dr. Porter P. Lowry (Missouri
Botanical Gardens); Prof. Jörg Ganzhorn, (Hamburg University); Prof. Alison Jolly
(Sussex University); Dr. Rob Brett (Fauna & Flora International); Dr. Paul Smith
(Royal Botanic Gardens, Kew); and Lisa Gaylord (Wildlife Conservation Society).
Further details are given in Appendix 1.
2.6 Rio Tinto’s NPI commitment
The goal of Rio Tinto’s biodiversity strategy is a ‘Net Positive Impact’ (NPI) on
biodiversity (Rio Tinto, 2004, 2008a). This means “minimising [sic.] the impacts
of our business and contributing to biodiversity conservation to ensure a region
ultimately benefits as a result of our presence” (Rio Tinto, 2008a). Rio Tinto’s
position on biodiversity is embedded in the company’s land use stewardship
standard (Rio Tinto, 2008b). The company’s environmental approach is described
in the policy document The Way We Work (Rio Tinto, 2009). The biodiversity
18 For some of these species, there has been insufficient study conducted outside the mining zone to
determine conclusively whether they are strictly endemic to the mining zone or more widespread.
19 More information on the project can be found in the Social and Environmental Impact
Assessment (Rio Tinto QM M, 2001) and at www.riotintomadagascar.com.
The Rio Tinto QMM
project will mine
ilmenite ore as a
mineral sand, to
provide titanium
dioxide to the world
market. Titanium is a
major ingredient in
steels and other
alloys; titanium
dioxide is the white
pigment found in most
paints and plastics.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 11
strategy was launched in 2004 at the IUCN World Conservation Congress in
Bangkok. Since then, Rio Tinto’s Chief Executive Officer, Tom Albanese, has
reaffirmed the NPI policy in a number of subsequent forums including the 2008
IUCN World Conservation Congress in Barcelona. Rio Tinto’s position statement
and guiding principles on biodiversity are presented in Boxes 2 and 3.
In simple terms, the NPI goal means ensuring that Rio Tinto’s actions have positive
effects on biodiversity that not only balance but are broadly accepted to outweigh
the inevitable negative effects of the physical disturbances and impacts associated
with mining and mineral processing. The company proposes to achieve this by:
• Avoiding unacceptable impacts on ecosystems.
• Reducing the impacts that may occur.
• Restoring impacted ecosystems.
• Compensating for residual impacts with offsets.
• Seeking additional opportunities to contribute to local conservation.
A key facilitator of the commitment to NPI is the development of comprehensive,
simple, scientifically sound metrics to quantify losses and gains. The lack of such
metrics has been a major reason for the lack of investment and enthusiasm by
developers (public and private alike) to attempt measurement and full mitigation
of biodiversity impacts (ten Kate et al., 2004).
Box 1: Rio Tinto’s position statement on biodiversity (from Rio Tinto, 2008a)
Rio Tinto recognizes that conservation and responsible management of biodiversity
are important business and societal issues. Our goal is to have a net positive
impact on biodiversity.
We are committed to the integration of biodiversity conservation considerations
into environmental and social decision making in the search for sustainable
development outcomes. We recognize that this might mean that we do not
proceed in some cases.
We want to be biodiversity leaders within the mining industry, for the competitive
advantage and reputational benefit this provides. Our performance on biodiversity
conservation and management issues will create benefits for our business.
We are committed to:
• The identification of biodiversity values impacted by our activities.
• The prevention, minimization, and mitigation of biodiversity risks
throughout the business cycle.
• Responsible stewardship of the land we manage.
• The identification and pursuit of biodiversity conservation opportunities.
• The involvement of communities and other constituencies in our
management of biodiversity issues.
The NPI goal means
ensuring that Rio
Tinto’s actions have
positive effects on
biodiversity that not
only balance but are
broadly accepted to
outweigh the
inevitable negative
effects of the physical
disturbances and
impacts associated
with mining and
mineral processing.
Box 1: Rio Tinto’s position statement on biodiversity (from Rio Tinto, 2008a)
12 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
The mitigation
hierarchy is a
conceptual framework
for thinking about
biodiversity risks and
opportunities and
developing
appropriate responses.
Box 2: Rio Tinto’s guiding principles on biodiversity (from Rio Tinto, 2008a)
• Our goal is to have a net positive impact on biodiversity by minimizing the
negative impacts of our activities and by making appropriate contributions
to conservation in the regions in which we operate.
• We are committed to the conservation of threatened and endemic species
and high priority conservation areas, and support local, national and global
conservation initiatives.
• We will seek equity and the reconciliation of differing perspectives and
ideals in biodiversity decisions and actions.
• We will enhance biodiversity outcomes through consultation, constructive
relationships, and partnerships with key stakeholders.
• We will integrate the identification, evaluation, and management of
biodiversity issues into the planning, decision making, and reporting
processes throughout the business cycle.
• We will apply appropriate expertise and resources to biodiversity issues,
building internal and external capacity where necessary.
• Subject to appropriate consent, we promote the collection, analysis, and
dissemination of biodiversity information and knowledge.
2.7 The mitigation hierarchy (avoidance,
minimization, restoration and rehabilitation,
and biodiversity offsets)
The mitigation hierarchy (Figure 2) is a conceptual framework for thinking
about biodiversity risks and opportunities and developing appropriate responses.
Variations on the mitigation hierarchy were first developed around 2004 by Rio
Tinto and others. It is now a well established model for private sector biodiversity
management and conservation and has been adopted by a number of government
initiatives and private sector organizations (McKenney and Kiesecker, 2010;
TEEB, 2010). Proper use of the mitigation hierarchy means one must first seek
to avoid impacts, then minimize, then restore, and finally only use offsets as
an option to compensate for the residual impacts after all other options have
been exercised (ten Kate et al., 2004; Rio Tinto, 2008a). The meaning, scope
and use of stages in the mitigation hierarchy are summarized in this section.
Box 2: Rio Tinto’s guiding principles on biodiversity (from Rio Tinto, 2008a)
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 13
Net positive impact
- Biodiversity values +
Additional
conservation
actions
Biodiversity
impact
Biodiversity
impact
Biodiversity
impact
Biodiversity
impact
Rehabilitation
MinimizationMinimization Minimization Minimization
AvoidanceAvoidanceAvoidance Avoidance Avoidance
Rehabilitation Rehabilitation
Offset Offset
Residual impact
Figure 2.
The mitigation hierarchy. From Rio Tinto and Biodiversity: achieving
results on the ground (Rio Tinto, 2008a).
Avoidance
Rio Tinto (2008a) defines ‘avoidance’ as activities that change the scope of
impacts (reducing them, moving them, or avoiding them completely). Avoidance
changes or stops actions before they take place, preventing their expected impacts
on biodiversity. It involves a decision to change the expected or normal course
of action. A real-world example can be found at Rio Tinto Simandou’s iron ore
project in southeastern Guinea where stockpiles and waste dumps have been
relocated to avoid impacts on tropical forest; they are instead mainly being located
in areas of relatively degraded savannah (a common habitat type in the region).
The biggest opportunity for avoidance is during project developmentit is often
possible to implement relatively cheap avoidance measures that significantly
reduce impacts on biodiversity, thereby reducing future costs of restoration,
offsets and closure.
Minimization
‘Minimization’ reduces the severity of impacts on biodiversity that result from
mining and associated activities20 already underway. These actions reduce the
likelihood or magnitude of biodiversity impacts, but cannot completely prevent
them. It can sometimes be difficult to distinguish between avoidance and
minimization because some actions have aspects of both. Improvements to
the water quality treatment of outflows from mining areas, thereby reducing
impacts on aquatic systems is a good example of minimization, while routing
water outflows away from biodiversity-sensitive areas would qualify as avoidance.
20 ‘Mining and associated activities’ include all activities required to find, mine and process
minerals, at any stage of the mine life cycle from exploration to closure.
The biggest opportunity
for avoidance is during
project developmentit
is often possible to
implement relatively
cheap avoidance
measures that
significantly reduce
impacts on biodiversity,
thereby reducing future
costs of restoration,
offsets and closure.
14 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Rehabilitation and restoration
‘Rehabilitation’ involves the preparation of safe and stable landforms on sites
that have been disturbed by mining and associated activities, followed by re-
vegetation with the aim of establishing a specific habitat type. Rehabilitation is
important for improving basic ecosystem functions such as erosion control and
water quality regulation. ‘Restoration’ is a term generally used where the aim
is to recreate a habitat type similar to the original vegetation type, including
the targeting of some specific biodiversity features such as rare species.21 The
re-establishment of dune forest at Richard’s Bay Minerals on recreated sand
dune systems (following ilmenite sand mining) is an example of attempted
restoration within Rio Tinto’s portfolio (van Aarde et al., 1996). For the purposes
of NPI calculations, typically restoration can count towards achieving NPI22 but
rehabilitation cannot.
Offsets
Rio Tinto is committed to achieving a Net Positive Impact on biodiversity. Even
with the best possible mitigation measures in place at business units, mining and
associated activities will result in some level of residual impacts on biodiversity.
Consequently biodiversity offsets are neededconservation activities in the
wider region that result in measurable biodiversity gains to compensate for
these residual losses, resulting in a Net Positive Impact at the regional level.
Offsets are not employed in place of appropriate on-site avoidance and
minimization measures, but rather seek exclusively to address the residual loss
after mitigation. Offsets can achieve biodiversity ‘gains’ in two ways. They may
reduce existing pressures and therefore losses to biodiversity (e.g. reducing
background deforestation rates)this is known as an ‘averted loss offset’.
Alternatively, they may directly enhance the state of biodiversity (such as through
species re-introductions or habitat restoration).
Additional conservation actions
‘Additional conservation actions’ include a broad range of activities that are
intended to benefit biodiversity, but where effects or outcomes are difficult
to quantify in terms of biodiversity gains. Examples include scientific research,
environmental education, and building capacity and expertise in conservation
organizations. Although the biodiversity outcomes of these actions are difficult
to measure, these kinds of intangible assets form an essential part of Rio Tinto’s
contribution to biodiversity conservation, often underpinning the success of
other mitigation actions, and are often some of the most highly valued by
interested stakeholders.
21 E.g. see Society for Ecological Restoration International Science & Policy Working Group (2004 ).
22 Partially successful restoration, or restoration that is in pro gress but has not yet reached its
end goal, will be accounted for pro rata based on the extent to which biodiversity value has
been restored. This can be taken into account using the Quality Hectares and Units of Global
Distribution metrics. Restoration is very seldom (or never) ‘100% success ful’ in returning an area to
a facsimile of it s previous ‘natural’ state (or to the appropriate defined benchmark). Nevertheless,
biodiversity gains from restoration attempts can be quantified.
Offsets are not
employed in place of
appropriate on-site
avoidance and
minimization
measures, but rather
seek exclusively to
address the residual
loss after mitigation.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 15
3 METHODS
To obtain the information needed to determine if Rio Tinto QMM is on track to
achieve NPI, the following steps were followed:
1. Identify and prioritize biodiversity features for inclusion in NPI accounting.
2. Decide which metrics to use.
3. Select the counterfactual scenario(s) against which to measure losses
and gains.
4. Quantify biodiversity losses likely to be caused by Rio Tinto QMM in the
periods 2004–2015 and 2004–2065.
5. Quantify biodiversity gains likely to be caused by Rio Tinto QMM in the
periods 2004–2015 and 2004–2065.
Each of these five steps is summarized below.
3.1 Identify and prioritize biodiversity features for
inclusion in NPI accounting
Biodiversity is complex and can be measured at many levels, but lacks a single
uniform and globally fungible metric (in contrast with carbon, for example).
In addressing this issue an attempt was made to identify metrics that were
both practical to measure and reflective of the impacts associated with the
development of the Rio Tinto QMM mining operation. This report therefore
only considers losses and gains in terrestrial systems and in intrinsic/existence
values of biodiversity. For the purposes of this study, this effectively means
consideration of natural habitats and species.
Losses and gains in aquatic systems and in service values of biodiversity (biodiversity-
based ecosystem services, livelihoods and cultural values) have been covered in
previous reports and discussion papers, including Rio Tinto (2008c) and Rio Tinto
QMM (2001, 2008); and several aquatic studies including Jacques Whitford Inc.
(2007). Aquatic systems and service values are not considered further in the
quantitative analysis presented here. Mitigation measures for these biodiversity
features are detailed in the Biodiversity Action Plan (Rio Tinto QMM, 2010).
Potential losses and gains were measured for the following biodiversity features:
• Habitats—all forest; littoral forest and its sub-types (including Fort
Dauphin-type littoral forest; losses and gains were measured individually
for Mandena, Petriky and Ste Luce).
• Species—all High Priority terrestrial species listed in the BAP (restricted-
range23 and/or highly threatened24 vertebrates, invertebrates and plants).
23 ‘Restricted-range’ here includes site endemics and near-endemics, and possible site endemics.
24 ‘Highly Threatened’ here includes species assessed as Critically Endangered or Endangered on
the IUCN Red List.
Biodiversity is complex
and can be measured
at many levels, but
lacks a single uniform
and globally fungible
metric (in contrast with
carbon, for example).
16 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
For further information on how biodiversity features were identified and prioritized,
refer to the QMM Biodiversity Action Plan (Rio Tinto QMM, 2010).
3.2 Decide which metrics to use
Two metrics (or currencies) were used—Quality Hectares (QH) and Units of
Global Distribution (UD). Quality Hectares are Rio Tinto’s current metric for
tracking progress towards the NPI target at the global and site levels. A wide
range of biodiversity values, including threatened species, rare habitats or
non-timber forest products, may be expressed in terms of their quantity and
quality. For example, 100 ha of forest in pristine condition would count as 100
Quality Hectares (100 ha × 100% quality = 100 QH), whereas 100 ha of fairly
degraded forest at 40% ‘optimum quality’ would be expressed as 40 Quality
Hectares (100 ha × 40% quality = 40 QH).
Units of Global Distribution are a novel metric, developed for this analysis, but
conceptually related to Quality Hectares. A Unit of Global Distribution is equivalent
to 1% of a species’ global population25 (or 1% of its global distribution,26 in the
event that population data are unavailable).27 Units of Global Distribution are
calculated as follows: if a species has a global population of 1,000 individuals,
and 10 of those are killed, that would be a loss of 1% of the global population or
1 ‘Unit of Global Distribution’ (UD). Similarly, if a species has a global distribution
of 100 ha, and 1 ha of its distribution is lost as a result of habitat loss caused
by mining, that would be a loss of 1% of its global distribution or 1 ‘Unit of
Global Distribution’ (UD).
A detailed discussion of both metrics is provided in Appendix 2.
Losses and gains were measured in Quality Hectares for all habitats considered.
For species, losses and gains were measured in Units of Global Distribution. For a
very small number of High Priority species it was not possible to measure losses
and gains in UD as global range and/or population size could not be quantified.
For these species, losses and gains were simply measured in hectares.
The UD metric provides additional information which is useful for making
‘like-for-not-like’ biodiversity offset comparisons. Most frequently, biodiversity
offsetting involves ‘like-for-like’ or ‘like-for-better’ (also known as ‘trading up’)
exchanges. For example, a like-for-like offset would occur when a loss of an
area of a particular habitat type is offset by proportionately equal or greater
gains in area in an essentially identical habitat type. A like-for-better offset might
25 Calculated in number of mature individuals.
26 Calculated in hectares.
27 It should be noted that assuming such a link between distribution and population size may be
particularly problematic for wide-ranging and nomadic species (it goes against ecological theory
on population and range size). However, no such species are included in the present analysis.
A precedent for making such a link is given in the Key Biodiversity Area guidelines (Langhammer
et al., 2007, p.65).
Quality Hectares are
Rio Tinto’s current
metric for tracking
progress towards the
NPI target at the
global and site levels.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 17
entail offsetting an area of low-value habitat (e.g. a very common habitat
type harbouring no threatened species) with a greater area of higher-value
habitat (e.g. a rare habitat type harbouring several threatened species). The
Rio Tinto QMM project is particularly complex because it potentially involves
like-for-not-like offsets, where the offset has different values to the area lost,
but it is not always objectively possible to claim that the values of the offset
site are greater.28
The relative values of different habitat types or biodiversity features (e.g.
what constitutes ‘trading up’) are fundamentally societal and thus require
subjective judgements and stakeholder consultation. In the case of Rio Tinto
QMM, guidance would be given by a range of stakeholders including (but not
restricted to) the Biodiversity Committee.29
3.3 Select counterfactual scenario(s) against which
to measure losses and gains
When measuring losses and gains, a key factor to consider is the counterfactual
scenario (or baseline) against which any loss or gain is measured. For Rio Tinto
QMM, this is particularly significant because the project is located in an area
which has experienced significant deforestation since at least the 1950s when
the first aerial photographs of the region were taken (Du Puy and Moat, 1998;
Vincelette et al., 2007b).
Three counterfactual scenarios were considered in the present analysis:
• No mining and ‘no deforestation’.
• No mining and a ‘national average’ deforestation rate for all forest types
extrapolated from c.1990c.2000 of 0.9% per year (Table 4 in Harper et
al., 2007).30
28 For example, the Bemangidy offset is humid forest rather than littoral forest. QMM’s target is
‘Net Positive Impact on littoral forest’, so in this specific case Bemangidy is a kind of ‘insurance
policy’ against incomplete mitigation/of fset success elsewhere because it contains many of the
same species, but it is also a like-for-not-like offset.
29 Such guidance has not been needed at QM M to date, because the Biodiversity Committee set
a relatively strict ‘like-for-like’ target requiring that NPI is achieved for (i) littoral forest, and (ii)
all priorit y species, individually. This target is described here as ‘relatively strict’ rather than ‘very
strict ’ because it could be argued that, for example, because ‘Petriky-type littoral forest’ is distinct
from ‘Mandena-type littoral forest’ (Dumetz, 1999), NPI should be achieved at the level of these
forest sub-types rather than for littoral forest as a whole. However the recommendation of the
committee at the end of the May 2010 meeting was to measure NPI at the level of littoral forest
sensu lato.
30 Note that there is much evidence to show that historic deforestation rates do not necessarily
reflect future deforestation rates. A number of REDD (Reducing Emissions from Deforestation
and Degradation—another methodology where background deforestation rate is very impor tant)
project s around the world are moving away from using past baselines to predict future rates,
to methods that that look at population density, roads, etc. to predict future deforestation. The
issue of calculating background deforestation rates is one that would merit further consideration
in future, as discus sed in the ‘Lessons learned’ section.
18 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
• No mining and a calculated Fort Dauphin regional average deforestation
rate extrapolated from c.1995c.2005 of 3.89% per year. Regional average
deforestation rate was calculated for the period 1995–2005 based on
digitized aerial photographs using GIS31.
Figure 3 shows how forest cover would be predicted to change from 2004 to
2065 for Scenarios 1–3 in remaining forest at Mandena, Petriky and Ste Luce. In
2004, there were 2,289 ha of littoral forest on the mining leases (Mandena, Petriky
and Ste Luce), of which 1,665 ha fell under the mine path. Within the Avoidance
Zones (AZs), 624 ha of littoral forest are protected.
2004
2007
2010
2013
2016
2019
2022
2025
2028
2031
2034
2037
2040
2043
2046
2049
2052
2055
2058
2061
2064
2500
2000
1500
1000
500
0
0%
0.90%
3.89%
Year
Hectares of littoral forest in Fort Dauphin area
Figure 3.
Predicted changes in littoral forest cover on the Rio Tinto QMM mining
leases from 2004 to 2065, under three different annual deforestation
rate scenarios—0% (no deforestation from 2004 onwards), 0.9%
(Madagascar national average) and 3.89% (Fort Dauphin regional average).
The counterfactual scenario is a critical element of the NPI loss and gain calculations
as it determines the magnitude of loss for which Rio Tinto QMM is considered
to be responsible (Figure 4). Point ‘a’ in Figure 4 shows that if there had been
no mine and if deforestation had continued at the Fort Dauphin regional rate of
3.89%, by c.2035 there would have been less forest remaining than the 624 ha
that are currently protected in the Avoidance Zones.
31 Note that although mining did not star t until 2009, Rio Tinto QMM was active in the region during
this period and its activities (e.g. road construction) may have facilitated access and indirectly resulted
in elevated rates of forest loss.
The counterfactual
scenario is a critical
element of the NPI loss
and gain calculations
as it determines the
magnitude of loss for
which Rio Tinto QMM
is considered to be
responsible.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 19
2004
2007
2010
2013
2016
2019
2022
2025
2028
2031
2034
2037
2040
2043
2046
2049
2052
2055
2058
2061
2064
2500
2000
1500
1000
500
0
0%
0.90%
3.89%
size of AZs
a
xy
z
Year
Hectares of littoral forest in Fort Dauphin area
Figure 4.
The three counterfactual scenarios considered. Distance x shows
that, assuming no mitigation beyond Avoidance Zones (AZs), based
on Scenario 1 (0% annual deforestation), Rio Tinto QMM would be
responsible for the loss of 1,665 ha of forest by 2065, i.e. the whole
area of forest on the ilmenite deposit (2,289 ha) minus the 624 ha
protected within the AZs. Distance y shows that, based on Scenario
2 (0.9% annual deforestation: national average), Rio Tinto QMM
would be responsible for the loss of 695 ha by 2065. Distance z shows
that, based on Scenario 3 (3.89% annual deforestation: Fort Dauphin
regional average), Rio Tinto QMM would be responsible for a net gain
of 421 ha of forest by 2065 simply by putting in place the Avoidance
Zones: the AZs protect an area of 624 ha of forest, whereas only 203
ha would be left by 2065 if forest loss continued at a rate of 3.89%
per year. Point ‘a’ shows the point at which, under Scenario 3, net
impact becomes positive based on AZs alone.
We suggest that out of these three possible counterfactual scenarios, Scenario
2the conservative national deforestation rate of 0.9%is the most appropriate
baseline against which to measure NPI. To use a 0% deforestation rate baseline
would be highly unrealistic, given the very high rate of deforestation that has
occurred since 1950 and continues to occur in the local region. The 3.89%
regional deforestation rate was not selected, partly to be precautionary, and
partly because it is problematic to use a rate that may have been caused (at
least in part) by the mine. For example, it is possible that prospecting activities
20 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
undertaken by Rio Tinto QMM since the 1990s, including road construction,
have indirectly facilitated the loss and degradation of the remaining forest
(e.g. Ingram and Dawson, 2006; Virah-Sawmy and Ebeling, 2010). Based on
aerial photographs from 1950, 1974 and 1989 (before the arrival of Rio Tinto
QMM), the annual deforestation rate in the Fort Dauphin region was 1.1% per
year (Vincelette et al., 2007b). Thus, adopting the 1990–2000 national average
(0.9%) as a baseline is conservative in comparison to this historical rate.
Furthermore, there is a view amongst some stakeholders that, given the occurrence
of many site endemic and Critically Endangered plants and animals, it is possible
that the remaining forests would have been designated a Protected Area at
some point in the next decade or so.32 If a protected area covering the whole
of Mandena, Petriky and Ste Luce were to have been put in place in 2020 (and
if that protected area was 100% successful in stopping forest loss), the forest
cover would have stabilized at 1,213 ha (Figure 4), assuming a background rate
of 3.89% annual loss. By comparison with this alternative scenario, Rio Tinto
QMM would be considered responsible for the loss of 589 ha. Consequently
this scenario is somewhat less conservative than Scenario 2 (under which Rio
Tinto QMM is considered responsible for the loss of 695 ha). Following the same
logic but assuming a 0.9% rate of loss, Rio Tinto QMM would be considered
responsible for the loss of 1,356 ha. This is more conservative than Scenario 2,
although Rio Tinto QMM would still be predicted to achieve NPI on the littoral
forest under this counterfactual scenario.33
3.4 Quantify habitat losses for the periods
2004–2015 and 2004–2065
Losses caused by Rio Tinto QMM’s mining activities were measured (past
losses) and predicted (future losses) for littoral forest habitat, using the Quality
Hectares metric. Only primary impacts of Rio Tinto QMM’s mining operations
were quantified (e.g. loss of habitat directly caused by mining and associated
activities such as building roads and other infrastructure). Secondary impacts
(e.g. potential negative impacts caused by invasive alien species brought into
the region by mine transport, increased human pressure on ecosystems caused
by prospective in-migration) are difficult to measure and no attempt has been
made to formally quantify them here. In Rio Tinto QMM’s particular case,
because there is so little littoral forest left in the Fort Dauphin region, it is almost
all within Rio Tinto QMM’s direct influence (e.g. part of the Avoidance Zones,
the Ste Luce offset or the wider mining lease). Consequently Rio Tinto QMM
has more influence over what happens in the whole region’s littoral forest than
would be typical of a mining operation in another area. Rio Tinto QMM has
32 This view was e xpressed verbally by Porter P. Lowry at the QMM Biodiversity Committee meeting
in May 2010.
33 Under Scenario 2 (0.9% deforest ation), net impact by 2065 is +350 QH of littoral forest. Under
an alternative scenario of 0.9% deforestation per year to 2020, followed by Protected Area
designation for Mandena, Petriky and Ste Luce forests and no fur ther forest loss, the project would
be projected to have a Net Positive Impact of c.+190 QH by 2065.
Losses caused by Rio
Tinto QMM’s mining
activities were
measured and predicted
for littoral forest
habitat, using the
Quality Hectares metric.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 21
programmes to monitor and mitigate secondary impacts such as those described
above; mitigation measures include those described elsewhere in this report
such as planting fast-growing tree plantations to meet community needs for
wood and fuel and to relieve pressure on littoral forest. Indeed, the success of
Rio Tinto QMM’s Avoidance Zones and offsets rests on the ability of Rio Tinto
QMM (in partnership with local communities, government and other relevant
organizations) to address these secondary impacts; without doing so it would
be very difficult to slow or halt deforestation.
Losses were estimated by mapping forest extent and assessing forest condition
at all three mining leases, and overlaying these maps with the predicted dredge
path and any other mining infrastructure using GIS. Because no littoral forest
condition index existed in Madagascar, Rio Tinto QMM developed a five-
category scale of forest condition, based on a range of habitat structure variables
that were measured in the field, particularly canopy cover (Vincelette et al.,
2007b). The five categories range from ‘very good’ to ‘extreme deterioration’
(Figure 5). The methodology for assessing forest condition is summarized in
the following section (a full description can be found in Vincelette et al., 2007b
and Rabenantoandro et al., 2007).
For Scenario 2 (no mining and national average deforestation rate of 0.9% per
year) and Scenario 3 (no mining and a Fort Dauphin average deforestation rate
of 3.89% per year), losses were adjusted to take into account the amount of
forest that would have been remaining by 2065 under these counterfactuals.
For Scenario 1 (0% annual deforestation), losses were not adjusted.
3.4.1 Assessing forest condition
The forest condition assessment method involves mapping all remaining littoral
forest blocks (based on the interpretation of the most recent aerial photographs
or satellite images) and establishing transects to cover each forest block with
a 50 × 50 m grid. Sample points are established at the grid intersections. The
following data are obtained at each sampling position within the grid: general
condition of the forest; signs of cutting (stumps); openings; agricultural areas;
fires; and observations of the vertical structure of the forest canopy level (upper,
intermediary, or lower). Finally, the field observer evaluates percentage canopy
cover at the sampling position.
There is a progressive decrease in canopy height and tree diameter at breast
height (dbh) from Ste Luce to Mandena to Petriky; this is a natural feature of
these areas that relates to climatic differences (e.g. Petriky sees markedly less
rainfall than Ste Luce; Rabenantoandro, et al. 2007). For example, in the study
plots sampled by Rabenantoandro et al. (2007), canopy height decreased from
a mean of 14.7 m in Ste Luce to 4.4 m in Petriky. Consequently it was necessary
to calibrate the method to reflect these intrinsic differences in stature, and so
the reference (‘100% optimal quality’) was different for each of the three sites.
22 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
The forest condition assessment was first carried out in 1998; the method used in
this assessment was verified by Missouri Botanical Garden (MBG), Royal Botanic
Gardens (RBG Kew), and FOFIFA
34
(Lowry et al., 2001). The forest condition
assessment was updated and improved based on new field data and Quickbird
images obtained in 2005 and by adding information from other studies on
the level of deterioration of a given block, dendrometric criteria, and floristic
composition (Henderson, 1999; Ingram and Dawson, 2005, 2006; Ingram et
al., 2005a, 2005b).
The analysis presented in the present report is based on the updated 2005
habitat condition assessment. Figure 5 shows the results of the 2005 forest
condition assessment.35
3.5 Quantify habitat gains for the periods
2004–2015 and 2004–2065
Measurable biodiversity gains can be generated either by increasing quality or
quantity (or both) of a given biodiversity value, for example a habitat type. The
key factor in each case is additionality—there must be a measurable increase
in quality or quantity that can reasonably be attributed to actions taken by Rio
Tinto QMM.
Three types of biodiversity gains are considered in the present analysisquality
gains and averted deforestation in the Avoidance Zones, restoration on
post-mining land, and averted deforestation at the biodiversity offset sites.
3.5.1 Restoration
At Rio Tinto QMM, restoration entails the replacement of natural habitat
surrogates following the completion of the mining process. Strictly speaking,
restoration on post-mining land does not represent a true gain in biodiversity
value, but instead reduces loss of biodiversity compared to a scenario where no
post-mining restoration was to take place, or a monoculture habitat was returned
post-mining. However for the sake of simplifying the loss-gain calculation we
treat post-mining restoration as a biodiversity gain.
Restoration gains were calculated in terms of Quality Hectares (area × quality),
assuming that by 2065 habitat will have been restored to 35% of optimal
quality at Mandena, 25% of optimal quality at Ste Luce and 20% of optimal
34 The National Centre for Applied Research into Rural Development, w ww.fofifa.mg.
35 Note that, in an earlier draft version of this analysis, a map showing forest cover and condition
in 1998 was erroneously included. However, the analyses presented in this report (and in previous
draft versions of the report) are based on forest condition assessment data from 2005.
The forest condition
assessment was first
carried out in 1998; the
method used in this
assessment was
verified by Missouri
Botanical Garden,
Royal Botanic Gardens,
Kew and FOFIFA.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 23
quality at Petriky.
36
This was estimated with reference to the littoral forest quality
measures developed by Vincelette et al. (2007b). Predictions of future forest
quality were made in consultation with botanical experts from RBG Kew and
Missouri Botanical Garden, along with Rio Tinto QMM’s in-house botanists. The
predictions took into account experience from restoration field trials carried out
since 1999 (Vincelette et al., 2007d).
These quality estimates are conservative, taking into account some uncertainties
that exist around restoration of Madagascan littoral forest. They are based on
the assumption that restoration at Mandena will start in c.2020, at Petriky in
c.2030 and Ste Luce in c.2035 (this is possible because Rio Tinto QMM is a
dredge mine; restoration can start on post-mining areas while mining is still
ongoing elsewhere on the site). They are also based on the assumption that
restoration at Petriky will be more difficult and quality gains will accrue more
slowly (this is the advice of botanical experts; Petriky is significantly more arid
and trees establish with greater difficulty and grow more slowly).
36 These assumptions about forest quality in 2065 were based on discussion with M. Vincelette,
J. Rabenantoandro and F. Randriatafika (who have been running littoral forest restoration trials
for 10+ years ), experience of success to date with the field trials, and consultation with botanical
experts on the biodiversity committee (Porter P. Lowry and Paul Smith ). The intention is to err on
the side of caution as littoral forest restoration has never be en attempted before.
24 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Figure 5.
Littoral forest extent and condition at Rio Tinto QMM in 2005.37
37 Note that, in an earlier draft version of this analysis, a map labelled as showing forest cover and
condition in 1998 was erroneously includ ed. However, the analyses presented in this repor t (and
in previous draft versions of the report) are based on the forest condition assessment data for
2005, which according to Vincelette et al. (2007) take into account the work of Henderson (1999) ,
Ingram and Dawson (2005, 2006), and Ingram et al. (2005a, 2005b).
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 25
3.5.2 Quality gains in the Avoidance Zones
The Avoidance Zones reduce the amount of habitat lost, by protecting areas of
forest on top of the deposit that would otherwise have been cleared for mining.
However, if Rio Tinto QMM carries out conservation activities that improve the
quality of the littoral forest in the AZs, it is valid to count this as a biodiversity gain.
Although the Avoidance Zones were situated to protect the best-quality habitat
remaining at Mandena, Petriky and Ste Luce, the habitat quality at each AZ site
varies considerably and is often far from pristine. In 2004, average quality scores
were 0.57 for forest fragments in Mandena AZ, 0.39 for Petriky AZ and 0.80 for
Ste Luce AZ (Figure 5). Consequently there is significant scope for the quality to
be improved, generating gains in QH. This will be achieved through appropriate
habitat management and by relieving pressure from local human populations by
inter alia providing plantations38 of fast-growing non-native39 trees outside the
Avoidance Zones for charcoal production and providing alternative livelihoods.
Quality scores were predicted to increase by 0.1 every 15 years in the Mandena
and Ste Luce AZs and by 0.05 every 15 years at Petriky.40, 41
3.5.3 Averted deforestation gains (long-term protection
of Avoidance Zones and biodiversity offsets)
Averted deforestation (also known as ‘averted loss’) generates biodiversity
gains both at Rio Tinto QMM’s biodiversity offset sites (Mahabo, Bemangidy,
Ste Luce Forests) and in the Avoidance Zones (Mandena, Ste Luce and Petriky
AZs). Mahabo, Ste Luce, Mandena and Petriky all contain littoral forest.
It is important that ‘averted deforestation’ is not confused with ‘avoidance’.
The Avoidance Zones (AZs) are, as their name implies, primarily an avoidance
measurethey represent a significant area of the deposit (c.8% of total ilmenite,
and 27% of remaining forest on the deposit) foregone in order to protect
habitat and in particular to save certain locally endemic species from extinction.
38 These plantations will principally be located in areas that have been cleared for mining, af ter
the dredge has passed (dredge mining moves steadily through the landscape, so such areas will
be available from an early stage in the project). We say ‘principally’ rather than ‘entirely’ because
some plantation has been established already as there is a pressing need among local communities
for wood; these plantations were sited taking into account the conservation value of the existing
landscape as well as human needs. There has been and will be no additional clearance of lit toral
forest to site these plantations.
39 Both native and non-native tree species (including littoral forest species, species from elsewhere
in Madagascar, and exotic species that were already found in the Fort Dauphin area) were included
in rehabilitation trials (Rarivoson and Mara, 2007; Vincelette et al., 2007), but ultimately the
decision was made to use non-natives.
40 The rates of improvement in quality score were estimated based on discussion with M.
Vincelette, J. Rabenantoandro and F. Randriat afika, and informed by progress with restoration
trials to date. Restoration trials star ted in earnest in 1999 (field observations and less formal
experiments related to littoral forest restoration had been ongoing since 1992), details can be
found in Vincelette et al. (2007d). These estimates were discussed and agreed with the Biodiversity
Committee (in particular Paul Smith [ RBG Kew] and Porter P. Lowry [ MBG]) at a workshop in Fort
Dauphin, Madagascar, in May 2010.
41 In comment s on an earlier draft of this analysis, Paul Smith ( RBG Kew) noted that “This is quite
conservative. But better to err on the side of caution.”
The Avoidance Zones
reduce the amount of
habitat lost, by
protecting areas of
forest on top of the
deposit that would
otherwise have been
cleared for mining.
26 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
However, had Rio Tinto QMM not taken active steps to protect the AZs, the
forest within them would have continued to decline at the same rapid rate as
before due to pressure from local human populations, particularly for charcoal
production. Consequently, over time, this averted deforestation is effectively an
additional measurable gain (in the two scenarios where a shifting baseline of
0.9% or 3.89% background annual deforestation rate is used; in the scenario
of 0% background deforestation on the mining leases there are no gains from
averted loss in the AZs).
The background rate of deforestation in the project area and surrounding region is
high. By implementing active conservation measures at offset sites and providing
alternative livelihoods to local communities, this high rate of deforestation can
be decreased, and consequently the area of forest remaining after one year (or
50 years) is greater than would have been the case if the offset sites had not
been brought under conservation management.
Gains accruing through averted deforestation were estimated based on the
assumption that the background deforestation rate will be reduced by 50%
through Rio Tinto QMM’s conservation activities in the period 2004–2065. For
example, in the case of a 1,000 ha forest that had been declining at a rate of
2% per year prior to 2004, one could assume an averted loss of 10 ha in the
first year (1000 ha × (0.02/2)).
Gains were calculated in an analogous way to compound interest on a bank
account, as follows.
Gains from averted loss were calculated as:
G = [x × (1-0.5y)z] - [x × (1-y)z]
Where G = gains; x = QH at a site; y = background deforestation rate and
z = years of intervention.
For example, from 2004–2015 (11 years of intervention), gains would be:
G = [x × (1-0.5y)11] - [x × (1-y)11]
However, if at one site interventions were only started in 2011, then from
2004–2015 there would be eight years of ‘business as usual’ and three years
of conservation intervention, and so gains would be:
G = [x × (1-y)8 × (1-0.5y)3] - [x × (1-y)11]
By implementing
active conservation
measures at offset sites
and providing
alternative livelihoods
to local communities,
this high rate of
deforestation can
be decreased.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 27
For the Avoidance Zones, which are under Rio Tinto QMM’s direct management
and subject to intensive management, it was estimated that the deforestation
rate would be reduced by 100%i.e. no further area would be lost. For the Ste
Luce Forests offset, which is under Rio Tinto QMM’s direct influence (unlike the
other offset sites), it was estimated that deforestation could be reduced by 75%.
3.5.4 Credit claims
A number of Rio Tinto QMM’s offset sites are co-funded by other organizations.
To ensure that credit is apportioned appropriately, the following rule was used in
these analyses when calculating biodiversity gains attributable to Rio Tinto QMM:
If Rio Tinto QMM leads on a particular project and starts it up and maintains
it, 100% of the resulting gains can be claimed. However, if Rio Tinto QMM
joins and co-funds a pre-existing project being led by another organization,
benefits are only attributed on a pro rata basis proportionate to investment.
This rule attributes gain in proportion to investment, whilst also providing an
incentive to make the first move. Following this rule Rio Tinto QMM is able to
claim 100% of gains at all offset sites except for Mahabo, where conservation
measures had been started with funding from Missouri Botanical Gardens
prior to the involvement of Rio Tinto QMM. In the case of Mahabo, annual
management costs are c.US$65,000, of which 63% is provided by Rio Tinto
QMMconsequently 63% of the annual biodiversity gains from the site can
be attributed to Rio Tinto QMM (assuming the same proportion of investment
continues in future).
3.6 Calculating species losses and gains for the
periods 2004–2015 and 2004–2065
For species included in the analysis (High Priority species as defined in the
Biodiversity Action Plan), losses and gains were calculated in Units of Global
Distribution (UD). For most species, losses and gains were initially calculated in
hectares. Each High Priority species was coded by occurrence at the following
sites: Mandena, Petriky, Ste Luce Deposit, Ste Luce Avoidance Zone, Ste Luce
Offsets (Ste Luce was sub-divided in this way because it is a larger site, and some
species occur in e.g. the Ste Luce Offsets but not the other two sub-sites), Mahabo
and Bemangidy. Area of distribution on the mining leases, and concomitant
losses, were calculated assuming that a species was found throughout the
whole surface area of forest at any site at which it occurred. Custom estimates
of predicted loss, based on detailed field mapping and population estimates,
were made for four locally endemic plant species at Petriky that were known
to have particularly patchy or restricted distributions, such that this assumption
For species included
in the analysis, losses
and gains were
calculated in Units of
Global Distribution.
28 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
would not be valid.
42
Custom estimates of predicted loss were also made for
the Critically Endangered gecko Phelsuma antanosy.
To convert losses and gains in hectares to Units of Global Distribution, the
total global Area of Occupancy (AOO) was estimated for all species. Different
methods were used to estimate global AOO depending on the taxonomic group
under consideration and the information available. For locally endemic plants,
global distribution area was calculated based on the total area of known sites
for most species43 and custom estimates for certain Petriky species. For most
birds, mammals, and amphibians, Extent of Occurrence (EOO) was measured
based on the polygon area of GIS distribution maps from the IUCN Red List of
Threatened Species (IUCN, 2009), and AOO was inferred as 10% of EOO (this is
a very rough approximation that will not hold in all cases; the 10% relationship
between AOO and EOO is implied by the thresholds set for Criterion B of the
IUCN Red List). For some birds, mammals and amphibians, a more detailed
estimate of AOO is given in the IUCN Red List (2009); this was used where
available. For reptiles, estimates were based on draft maps provided by R. Jenkins
and prepared for the January 2011 IUCN Global Reptile Assessment workshop
held in Madagascar. For a very small number of High Priority species, it was not
possible to measure losses and gains in UD as global range and/or population
size could not be quantified based on existing data. For these species, losses
and gains were simply measured in hectares.
42 Eligmocarpus cynometroides, Eulophia filifolia, Myrtus madagascariensis, and Peponium
poissonii. Note that E. filifolia may have recently been found at Mahabo (C. Birkinshaw pers.
comm. 2010), but is precautionarily retained as a local endemic in the present analysis.
43 This is arguably too generic a use of the term ‘area of occupancy’, which has a very specific
definition in the IUCN Red List guidelines, and a different term might be more appropriate.
The intention here is to provide an approximation, albeit rough, of the size of the distribution
area where a species is actually found; it is well known that broad measures such as extent of
occurrence t ypically overestimate (sometimes vas tly so) the area occupied by each species (J etz,
Sekercioglu and Watson, 2008; Ro drigues, 2011).
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 29
4 RESULTS
4.1 Is Rio Tinto QMM on track to achieve a Net
Positive Impact on biodiversity for the period
2004–2065?
4.1.1 Summary
Biodiversity losses and gains were calculated in two ways: using Quality Hectares
as a metric (for forest, littoral forest and littoral forest sub-types); and using
Units of Global Distribution as a metric (for High Priority species, i.e. Highly
Threatened and restricted-range species).
Three offset sites were included in the quantitative analysisSte Luce Forests,
Mahabo and Bemangidy. Two of these sites (Ste Luce Forests and Mahabo)
already have active conservation programmes, the third (Bemangidy) has a draft
management plan and is awaiting confirmation as an offset from Rio Tinto
QMM and the Biodiversity Committee.
Two further sites (Ambatotsirongorongo and TGK I Direct Payments Project) have had
active conservation projects, but are currently under review regarding whether they
should be maintained as offset sites in the future. In the case of Ambatotsirongorongo,
work at the site will continue in future44, although this may be formally classed as an
additional conservation action rather than an offset within the Rio Tinto mitigation
hierarchy (in this case because of the high risk that even well-managed conservation
projects may not deliver measurable biodiversity gains because of the very great
pressure that the site is under – Ambatotsirongorongo is very small, fragmented,
and degraded as a result of pre-existing threats). Consequently, these two sites have
not been included as offset sites in the quantitative analysis presented here. Further
details on each of the offset sites are given in the discussion.
Losses and gains were therefore calculated based on the following assumptions:
• Rio Tinto QMM will maintain three offset sites: Mahabo, Bemangidy and
Ste Luce Forests.
• 225 ha of restoration will be carried out at Mandena, Petriky and Ste Luce,
respectively (775 ha in total).
• Avoidance Zones at Mandena, Petriky and Ste Luce will be maintained and
enhanced.45
44 Rio Tinto QMM is currently contributing to two projects at Ambatotsirongorongo : a natural
resources management project with UNDP and the local NGO FAFAFI; and a management plan
for the Critically Endangered gecko Phelsuma antanosy with Fauna & Flora International and the
national NGO Voakajy.
45 ‘Enhanced’ means enhanced in quality rather than increased in area. This may involve
e.g. enrichment planting of High Priority plant species, where this is deemed appropriate by
botanical experts.
Biodiversity losses
and gains were
calculated in two
ways: using Quality
Hectares and Units of
Global Distribution.
30 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Additionally, a series of drop-out analyses were carried out to examine the
impact on NPI of partial or total failure of restoration or offsets (e.g. removing
the offsets component from the analysis or greatly reducing the magnitude
of projected gains, and seeing whether NPI could still be reached based on
restoration and Avoidance Zone gains alone). Rio Tinto QMM does not envisage
that the restoration, Avoidance Zones or offset sites will fail, but it is important
to consider this possibility when forecasting to ensure that there is a sufficient
buffer to secure NPI even in the case of partial failure.
Based on the portfolio of offsets, restoration and Avoidance Zones described
above, for the period 2004 to 2065 (the latter being the current mine closure
date), Rio Tinto QMM is predicted to have a Net Positive Impact on the forests
in general, and on the littoral forest in particular.
Loss of littoral forest caused by direct impacts of mining is predicted to be -428
QH. Total gain of littoral forest is predicted to be +778 QH. Consequently net
impact is positive, at +350 QH, and the ratio of gain to loss (or compensation
to impact) is approximately 2:1. Considering all forest types, loss remains
constant at -428 QH; gain in all forest types (including Bemangidy humid forest)
is +1,679 QH. In this case the ratio of gain to loss (or compensation to impact)
is approximately 4:1. This information is particularly relevant given discussion
in the biodiversity offsetting community around multipliers (BBOP Multipliers
Consultation Working Group, 2008).
In terms of Units of Global Distribution, there is predicted to be a Net Positive
Impact on all High Priority plants (54/54 species) and most High Priority animals
(29/36) over the same time period.
Thus, if Net Positive Impact is to be achieved by 2065 overall, urgent research
and action are necessary to mitigate the residual impacts on the remaining seven
animal species, along with efforts to ensure the continued implementation of
current mitigation measures for the rest of the region’s species and habitats.
Table 1.
Predicted net impact of Rio Tinto QMM for the period 2004–2065,
based on Scenario 2 (0.9% annual deforestation rate, equivalent to
the Madagascar average).
2004-2065
Quality Hectares 1. All forest +1,251
2. Littoral forest +350
3. Fort Dauphin littoral forest
(including Mandena, Petriky, Ste
Luce; excluding Mahabo)
+216
Units of
Global Distribution
1. All High Priority species 83/90 positive
2. Priority plants only 54/54 positive
3. Priority animals only 29/36 positive
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 31
4.1.2 Impacts on habitats—Quality Hectares
4.1.2.1 Results for 2004–2065
The results of this NPI forecast show that, in terms of Quality Hectares, Rio
Tinto QMM is predicted to have a Net Positive Impact in the period 2004–2065,
based on the conservative national average deforestation rate of 0.9% per year.
Net Impact is predicted to be positive for littoral forest (Mandena, Petriky, Ste
Luce and Mahabo; Table 2, Figure 6), and for all forest types combined (i.e.
including the humid forests of the Bemangidy offset) (Table 2, Appendix 3).
Net impact on littoral forest is forecast to be +350 QH in 2065 (Table 2),
representing an increase of 13% by comparison with 2004, when there were
2,747 QH of littoral forest in Fort Dauphin (Mandena, Petriky and Ste Luce) and
Mahabo, collectively.
A similar result is seen if the analysis is simply based on hectares of littoral
forest rather than Quality Hectaresin this case, by 2065, the net impact is an
increase of 205 ha (an increase of 5% by comparison with 2004 forest cover,
which was 4,352 ha in Fort Dauphin and Mahabo).
As can be seen in Figure 6, a similar magnitude of gains in QH is generated by
restoration (collectively), averted loss in Avoidance Zones (collectively), quality
gains in the Avoidance Zones (collectively) and averted loss in the Ste Luce and
Mahabo offsets. Gains at Bemangidy are predicted to be of significantly greater
magnitude; however as this is humid forest it is not a like-for-like offset for
littoral forest (although there is significant overlap of species, c.50% for plants:
Rabenantoandro et al., 2007).
Based on the results per site presented in Table 2, it is possible to calculate net
position for any individual site or combination of sites. For example, Petriky on
its own (which is distinct from Mandena and Ste Luce, see Section 3.3.1.2),
is forecast to be at a net position of -9 QH in 2065. Looking at Mandena, Ste
Luce and Petriky together but excluding Mahabo (which may be relevant from
a conservation perspective as Dumetz, 1999, classified these three forests as
a unique type of littoral forest on sand), net position in 2065 is forecast to be
+216 QH.
32 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Table 2.
Losses and gains in QH predicted for 2004–2065 for each category of
loss and gain and by forest type, based on counterfactual Scenario 2
for the Fort Dauphin region (i.e., 0.9% annual deforestation rate,
equivalent to the Madagascar average).
Type of loss/gain Site
QH of
forest
lost
/ gained
Hectares
of forest
lost /
gained
LOSSES Mandena mining -23 -208
LOSSES Petriky mining -98 -425
LOSSES Ste Luce mining -307 -417
RESTORATION GAINS Mandena restoration 79 225
RESTORATION GAINS Petriky restoration 45 225
RESTORATION GAINS Ste Luce restoration 56 225
AVERTED LOSS IN AZs Mandena Avoidance 56 97
AVERTED LOSS IN AZs Petriky Avoidance 20 51
AVERTED LOSS IN AZs Ste Luce Avoidance 92 116
INCREASED QUALITY IN AZs Mandena Avoidance 92 0
INCREASED QUALITY IN AZs Petriky Avoidance 24 0
INCREASED QUALITY IN AZs Ste Luce Avoidance 56 0
AVERTED LOSS IN OFFSETS Ste Luce Offset 124 147
AVERTED LOSS IN OFFSETS Mahabo (littoral forest) 134 168
AVERTED LOSS IN OFFSETS Bemangidy (humid forest) 901 1001
Total all forest types 1251 1206
Total Fort Dauphin
littoral forest 216 37
Total littoral forest
habitat type (Mandena,
Petriky, Ste Luce
and Mahabo) 350 205
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 33
Avoidance Zones (AZs)
Averted loss in AZs
Increased AZ quality
Restoration
Offsets
Impacts
1800
1000
500
0
-500
-1000
Biodiversity impact
/ measured in QH
Figure 6.
Losses and gains of Quality Hectares of forest for the period 2004–
2065. Note that ‘offsets’ includes both the littoral forest offsets
of Ste Luce and Mahabo (+259 QH) and the humid forest offset of
Bemangidy (+901 QH).
4.1.2.2 Alternative scenarios
As explained in the discussion of the Methods, one assumption that makes
a significant difference to the outcome of the NPI analysis is the background
deforestation rate used for the Fort Dauphin region. The Madagascar national
average from c.1990c.2000 (0.9% annual deforestation) was selected as the
most appropriate precautionary baseline against which to forecast and measure
Rio Tinto QMM’s performance against the goal of reaching NPI.
However, in order to fully understand the implications of selecting this baseline,
alternative baselines were analysed for comparison. Littoral forest QH was
calculated looking at three different scenarios: 0%, 0.9% (national average)
and 3.89% (Fort Dauphin regional average 1995–2005). The method used
was the same in each case (e.g. see Appendix 3; to recalculate NPI for different
scenarios, the background deforestation rate for Mandena, Petriky and Ste Luce
was changed; all other variables remained the same). Additionally, a scenario of
0.1% was calculated to clarify the relationship between NPI and deforestation
rate, which is non-linear (Figure 7).
One assumption
that makes a
significant difference
to the outcome of
the NPI analysis is
the background
deforestation rate
used for the Fort
Dauphin region.
Cumulative QH gains
Impacts
34 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Net impact (Quality Hectares)
Background deforestation rate (% per year)
1200
1000
800
600
400
200
0
-200
0-0.5 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Figure 7.
NPI forecast (net impact for 2004–2065) based on different presumed
background annual deforestation rates (0%, 0.1%, 0.9%, 3.89%) for
the Fort Dauphin region.
As can be seen in Figure 7, net impact from 2004 to 2065 is negative (-68
QH) if the background deforestation rate for the Fort Dauphin region is set at
zero. This represents a reduction in littoral forest QH of 2% by comparison with
2004 levels. However, as soon as it is assumed that even a small degree of forest
loss is likely to have occurred in the absence of Rio Tinto QMM between 2004
and 2065 (between 0.1% and 0.2%; the point shown just below the x axis on
Figure 7 is based on a background rate of 0.1%), net impact becomes positive.
At the 1995–2005 measured regional background rate (3.89%), the Rio Tinto
QMM project would have a major positive impact (952 QH, representing an
increase of 35% by comparison with the 2004 littoral forest QH).
4.1.2.3 Risk of failure of offsets or restoration
A series of drop-out analyses was carried out to examine the impact upon NPI
at 2065 of total or partial failure of offsets or restoration to deliver measurable
gains. This could happen, for example, if external factors outside Rio Tinto
QMM’s control caused the deforestation rates at the off-site offsets to remain
as high as before 2004 (e.g. major political or socio-economic problems that
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 35
rendered Rio Tinto QMM and its partners’ best conservation efforts ineffectual
in reducing deforestation). In the case of restoration, no other organization
has ever attempted to restore littoral forest and, although Rio Tinto QMM has
made significant investment in research and restoration trials over the past 10
years, with good progress, there is a risk that restoration may not deliver even
the conservatively estimated gains forecast by the present analysis. All drop-
out analyses were based on a precautionary shifting baseline (0.9% annual
deforestation rate, the Madagascar average).
If restoration completely fails, the net impact at 2065 is forecast to be +170
QH of littoral forest (a gain of 6% by comparison with 2004). If both off-site
offsets (Mahabo and Bemangidy) fail, the net impact at 2065 is forecast to be
+216 QH of littoral forest (a gain of 8% by comparison with 2004).
However, if both restoration and offsets fail completely (and only the Avoidance
Zones are successful), the project would have a net negative impact of -88 QH
(3% loss).
4.1.3 Impacts on species—Units of Global Distribution
4.1.3.1 Results for 2004–2065
Biodiversity losses and gains were also calculated in terms of Units of Global
Distribution (UD) for High Priority species. This is essentially an extension of the
Quality Hectares method that calibrates losses and gains in terms of % global
range/population size. All High Priority species analyses were based on the
Scenario 2 baseline (0.9% annual deforestation rate, the Madagascar average).
Of the 90 High Priority terrestrial species (54 plants, 26 invertebrates, 10
vertebrates; Appendices 4 and 5), 83 (92%) are forecast to show a Net Positive
Impact by 2065. For 59 of these, area-based calculations predict that NPI will
be reached. For a further 24 plant species, area-based calculations predict a
moderate negative impact (-1.3% to -17.9%) but it is predicted that NPI can
be reached by 2065 through enrichment of the AZs and restoration zones
(e.g. planting species at a somewhat greater density than they currently occur
in nature).46 As Rio Tinto QMM tracks its progress towards achieving NPI over
the coming years and decades, gains in priority plant species will be measured
in terms of UD based on population sizethis means that enrichment gains
can be accounted for.
46 Because some of the littoral forest at Mandena, Petriky and Ste Luce was already degraded
prior to the arrival of QMM, it is considered likely that some priority plant species ( especially late-
succession species requiring a closed canopy) currently occur at a lower density than they would
do in optimal conditions. Consequently ‘enrichment planting’ to increas e the density of priority
species in the AZs and restoration areas can be seen as restoring habitat to an optimal state, rather
than creating an artificial landscape. QMM is being advis ed in these matters by experts in in situ
and ex situ plant conservation from RBG Kew and MBG.
Of the 90 High
Priority terrestrial
species 83 are
forecast to show a
Net Positive Impact
by 2065.
36 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Seven animal species show residual negative impactsfour vertebrates (two frogs,
two reptiles, including the Critically Endangered gecko Phelsuma antanosy) show
residual negative impacts of up to 5.1%, and three invertebrates (all millipedes)
show residual negative impacts of up to 17.9% (Table 3). It is not known whether
enrichment or ex situ conservation would be feasible for these species. Research
is underway to investigate options for achieving NPI for Phelsuma antanosy.
In the majority of cases, achievement of NPI for individual High Priority species is
dependent upon the success of restoration efforts. Targeted restoration will be
needed as it cannot be guaranteed that species will naturally colonize the restoration
zones. At present, 27 of the 54 High Priority plant species are being propagated
in a nursery, and it is intended that all High Priority species will be included in the
near future; careful attention is needed in the next few years to check that all
High Priority plants can be successfully propagated and planted out (e.g. into the
Avoidance Zones; many are late-successional species and consequently cannot
be planted in restoration zones until many years from nowtrials in the AZs are
recommended, so that any potential problems can be identified early).
Ex situ conservation measures (seed banking, establishment of populations in
botanical gardens, etc.) are needed for all High Priority plant species. Currently
17 of the 54 High Priority plant species (31%) have been stored in seed banks;
the aim is that seeds of all High Priority plant species should be banked by 2015
(Rio Tinto QMM, 2010). Some ex situ actions such as seed banking serve as a
kind of ‘insurance policy’, others such as establishment of populations in botanic
gardens can produce measurable gains in UD when reintroduced into the wild.
These measures will help to ensure a net positive impact on all species, and in
particular those species for which the NPI forecast predicts residual losses in wild
populations in 2065. It is a general principle in conservation that ex situ conservation
should not replace in situ conservation, although it is often an essential part of
conservation strategies for very rare or threatened species. Because ex situ and
in situ biodiversity gains are not the same, Rio Tinto QMM will account for these
gains separately as it tracks its future performance towards NPI.
Further trials and monitoring are needed to determine whether High Priority
animals will naturally colonize restoration zones or whether active measures
(e.g. translocation or captive breeding and release) are needed and would be
effective. Phelsuma antanosy shows a potential residual loss of 5.1% of its
global population; further research is needed to determine the most appropriate
conservation measures for this species and Rio Tinto QMM are supporting this
research through a Phelsuma antanosy Management Plan project with Fauna
& Flora International and the national NGO Voakajy. It has been protected at
Ambatotsirongorongo forest through a joint Rio Tinto QMM-Wildlife Conservation
Society (WCS) initiative
47
however this site was already very small and degraded,
47 WCS are no longer as active at Ambatotsirongorongo as they previously were; however Rio Tinto
QMM retains engagement with the site in the form of two projec ts: a natural resources management
project with UNDP and the local NGO FAFAFI; and a management plan for the Critically Endangered
gecko Phelsuma antanosy with Fauna & Flora International and the national NGO Voakajy.
Ex situ conservation
measures such as
seed banking and
establishment of
populations in
botanical gardens are
needed for all High
Priority plant species.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 37
and further research is necessary to determine whether the local population
is likely to be viable in the long term. ‘Gains’ from Ambatotsirongorongo are
not included in the summary figures for Phelsuma antanosy presented here.
Table 3.
Animal species with a net negative impact at 2065.
Group Species IUCN Red
List category*
Net
impact (UD)
Amphibians Guibemantis (Mantidactylus)
bicalcaratus ?sp. nov. NE -1.6
Amphibians Madecassophryne truebae EN -0.3
Reptiles Pseudoxyrhopus kely EN -3.0
Reptiles Phelsuma antanosy CR -5.1
Giant pill-millipedes Zoosphaerium alluaudi NE -17.9
Giant pill-millipedes Sphaeromimus splendidus NE -12.7
Spirobolid millipedes Alluviobolus laticlavius NE -17.9
*NE = Not Evaluated, EN = Endangered, CR = Critically Endangered
For all High Priority species together (note that individual results per species
were presented at the beginning of this section), there is forecast to be a net
gain of +1,256 UD. If site-endemic and Critically Endangered plants and animals
that occur at the offset sites but not on the mining leases are also considered,
there are additional gains of +493 UD (+60.3UD for animals and +432.7UD for
plants; Tables 4 and 5). These comprise 19 plant species with gains of 10-25
UD and six vertebrate species with gains of 1-25 UD (Tables 4 and 5). Overall,
the project is thus forecast to result in a gain of like-for-like and like-for-not-
like High Priority species of c.+1,750 UD. The net losses and gains for all High
Priority species are presented here for illustrative purposes and to highlight the
additional benefits that the offset sites bring for High Priority species that are
not found within the Rio Tinto QMM mining area. They should not be taken to
imply that one species can be exchanged for another species.
If restoration fails completely to generate measurable gains in any of the High
Priority species, there would be a deficit of -1,426 UD. The like-for-not like gains
of +493 UD at Mahabo and Bemangidy provide a partial but insufficient buffer in
case of complete failure. By expanding the size of the offset at Bemangidy (e.g.
from the currently proposed 4,000 ha to c.10,000 ha) a more complete buffer
could be provided, although it should be borne in mind that this would not be
like-for-like. Ex situ conservation will also help to provide a backup and buffer
in case of failure. Rio Tinto QMM intends to make its restoration a success, but
it is important to consider the worst case scenario during planning.
38 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Table 4.
Like-for-not-like gains in site-endemic plant species that occur at
offset sites but not on the mining leases.
Site Species
Species
distribution
(ha)
Gains
2004–
2065 (ha)
Gains
2065 as %
global range
Bemangidy
Lowryanthus rubens Pruski, gen. et sp.
Nov., ined. 4,000 1,001 25.03
Bemangidy Gnidia razakamalalana Z.S. Rogers 4,000 1,001 25.03
Bemangidy Ixora bemangidiensis Guédès 4,000 1,001 25.03
Bemangidy
Micronychia bemangidiensis Randrian.
& Lowry 4,000 1,001 25.03
Bemangidy
Diospyros bemangidiensis G.E. Schatz
& Lowry, sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Diosypros “Sclerophylla group” sp.
14, ined. 4,000 1,001 25.03
Bemangidy
Hyperacanthus gereaui Rakotonas. &
A.P. Davis, sp. nov. inéd. 4,000 1,001 25.03
Bemangidy
Hyperacanthus rajeriarisoniae
Rakotonas. & A.P. Davis, sp. nov. inéd. 4,000 1,001 25.03
Bemangidy
Ivodia anosiensis Rabarimanarivo et al.,
sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Polyscias bemangidiensis Lowry & G.M.
Plunkett, sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Polyscias ericii Lowry & G.M. Plunkett,
sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Polyscias manonae Lowry & G.M.
Plunkett, sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Polyscias urceolata Lowry & G.M.
Plunkett, sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Schefflera bemangidiensis Lowry &
G.M. Plunkett, sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Schizolaena charlotteae Lowry et al.,
sp. nov. ined. 4,000 1,001 25.03
Bemangidy
Schrebera trifoliata C. Frasier & G.E.
Schatz, sp. nov., ined. 4,000 1,001 25.03
Mahabo Brackenridgia sp. nov. 1,565 168 10.74
Mahabo Cassinopsis sp. nov. 1,565 168 10.74
Mahabo Octolepis cf dioica 1,565 168 10.74
432.67
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 39
Table 5.
Like-for-not-like gains in site-endemic, Endangered and Critically
Endangered animal species that occur at offset sites but not on the
mining leases.
Site
Higher
taxon
Species
Priority
Strict endemic
Regional endemic
IUCN EN or CR?
Global distribution of
species (ha)
Total area of site (ha)
Condition
Total QH at site
Gains 2004-2065 (ha)
Gains 2065 as %
global range
Bemangidy Amphibia
Mantidactylus
aff.grandidieri High 14,000 4,000 0.9 3,600 1,001 25.03
Bemangidy Amphibia
Boehmantis
microtympanum High 1EN 50,000 4,000 0.9 3,600 1,001 2.00
Mahabo Amphibia
Heterixalus sp.
nov. High 11,565 1,565 0.8 1,252 166 10.74
Mahabo Reptilia
Phelsuma cf
quadriocellata High 11,565 1,565 0.8 1,252 168 10.74
Mahabo Reptilia Phelsuma sp. nov. High 11,565 1,565 0.8 1,252 168 10.74
Mahabo Mammalia
Eulemur
cinereiceps High EN 1,5613 1,565 0.8 1,252 168 1.08
60.33
40 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
4.2 Is Rio Tinto QMM on track to achieve a
Net Positive Impact on biodiversity for
the period 2004–2015?
4.2.1 Impacts on habitats—Quality Hectares
4.2.1.1 Results for 2004–2015
The QH analysis was also carried out for the period 2004–2015, as 2015 is the
first internal Rio Tinto benchmark date when all business units subject to the
NPI target must measure their progress towards meeting this goal.
Rio Tinto QMM is predicted to have a Net Positive Impact on littoral forest QH
in the period 2004–2015, based on the national average deforestation rate
of 0.9% per year. NPI is predicted to be positive for littoral forest (Mandena,
Petriky, Ste Luce and Mahabo), and for all forest types combined (i.e. including
the humid forests of the Bemangidy offset as well).
Net impact on littoral forest is forecast to be +48 QH in 2015, representing
an increase of 2% by comparison with 2004 levels. This comes primarily from
avoided deforestation, as restoration will only recently have started.
4.2.2 Impacts on species—Units of Global Distribution
The UD analysis indicates that only 32 of the 90 High Priority species are likely
to be at NPI by 2015, and some species show losses of up to 39 UD (two
millipedes, Sphaeromimus inexpectatus and Riotintobolus mandensis, and one
plant Eulophia palmicola show population reductions of this magnitude; all of
these are Mandena endemics). This is because there will already have been major
impacts by this date (particularly at Mandena), but restoration zones will not
yet have reached sufficient maturity to count towards NPI. Carefully targeted
restoration is critically important for minimizing residual losses and achieving NPI
for High Priority species. However even at these sites, where research and trials
of restoration techniques were instigated well in advance of the start of mining
operations, it still takes time for restoration to create functioning natural habitat
analogues. Furthermore, many of the High Priority species are late-successional
species and cannot be planted out in restoration zones until the restored forest
is reasonably mature. Consequently, even with the best possible conservation
action, it is likely that the majority of High Priority species will not reach NPI by
2015, based on predicted losses and gains in their wild populations.
Carefully targeted
restoration is
critically important
for minimizing
residual losses and
achieving NPI for
High Priority species.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 41
4.3 Is Rio Tinto QMM on track to achieve a Net
Positive Impact on biodiversity throughout the
lifecycle of the mine?
Temporal loss is recognized as a significant issue that needs to be tackled
by biodiversity offsetting and ‘no net loss’ or ‘net positive impact’ initiatives
(ten Kate et al., 2004; Burgin, 2008; Bekessy et al., 2010). It can be said to
occur when a loss occurs in advance of sufficient gains being accrued through
restoration, averted loss at offsets, or other measures. To determine whether Rio
Tinto QMM is forecast to be ‘net negative’ with respect to littoral forest habitat
at any point during the lifecycle of the mine, losses and gains were calculated
using the same methods and assumptions as above and displayed on a year-
by-year cumulative basis (Figure 8). To calculate losses on a year-by-year basis,
some additional assumptions had to be made. These were (i) that Mandena,
Ste Luce and Petriky would be mined sequentially (this is the current plan), (ii)
that a constant number of QH will be lost to mining every year of ‘active’ mine
life (during the last few years of mine life, the ‘closure phase’, there will be
no further mining activity; activities such as decommissioning and restoration
and rehabilitation are carried out during this phase); and (iii) that mining will
stop c.12 years before final closure (this is the current plan). Assumption (ii) is
simplistic, first because although the dredge moves steadily and more-or-less
constantly through the landscape, covering c.100 ha per year (Rio Tinto QMM,
2001), the forest cover is patchy both in terms of quantity and quality. Second,
the extent of forest QH at Ste Luce is greater than that at Petriky, so the timing
of losses depends on whether Ste Luce or Petriky is mined after Mandena.
However, despite these caveats, better predictive data are not yet available and it
was felt that a year-by-year cumulative analysis would provide useful additional
information. This analysis was carried out for forest habitats only, measured in
Quality Hectares, because it would have been very significantly more complex
to carry out an analogous analysis for High Priority species.
The results (Figure 8) suggest that there will be no temporal loss of littoral forest,
and that Rio Tinto QMM will remain net positive in terms of littoral forest QH
throughout mine life.
Temporal loss is
recognized as a
significant issue that
needs to be tackled by
biodiversity offsetting
and ‘no net loss’ or
‘net positive impact’
initiatives.
42 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
1800
1600
1400
1200
1000
800
600
400
200
0
Bemangidy offset
Littoral offsets
Restoration
Increased AZ quality
Averted loss in AZs
Impacts
Time (2004-2065)
Cumulative QH gains / impacts measured in QH
Figure 8.
Cumulative gains in littoral forest over time (measured in QH)
compared to cumulative impacts from mining (measured in QH).
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 43
5 CONCLUSIONS
AND NEXT STEPS
5.1 Summary
Detailed analyses based on the best available scientific information were carried
out to project biodiversity losses and gains over the life of the mine (2004–2065)
and to determine whether Rio Tinto QMM will have a Net Positive Impact at
closure. Biodiversity losses and gains were calculated in two ways: using Quality
Hectares as a metric (for forest habitats); and using Units of Global Distribution
as a metric (for High Priority species, e.g. threatened and restricted-range
species). Rio Tinto QMM’s Biodiversity Committee has approved the use of these
metrics. Results show that Rio Tinto QMM is anticipated to have a Net Positive
Impact on biodiversity, both in terms of Quality Hectares of littoral forest and
Units of Global Distribution for the majority (83/90) of High Priority species, if
the mitigation measures outlined in this report are successfully implemented.
The mitigation portfolio proposed by Rio Tinto QMM is as follows:
Offsets: at (i) Mahabo, (ii) Tsitongambarika III Bemangidy, (iii) Ste Luce Forests.
Ambatotsirongorongo and the Tsitongambarika (TGK I) Direct Payments project
are currently under review as formal offsets, so gains from these two sites are
not included in the quantitative analysis presented here. Each of these sites is
discussed in more detail later in the report.
Restoration: 225 ha of restoration each at Mandena, Petriky, Ste Luce. Restoration
zones will be located adjacent to the Avoidance Zones, to provide a buffer,
improve connectivity and facilitate natural regeneration and recolonization.
Avoidance Zones: at Mandena, Petriky, Ste Luce.
The project is predicted to have a Net Positive Impact on littoral forest in 2065
(+350 QH), and a significant Net Positive Impact on regional forest types,
including Bemangidy humid forest (+1,251 QH).
The project is predicted to have a Net Positive Impact on 83/90 High Priority
species, comprising a total of +1,256 UD (including like-for-like species only)
and +c.1,750 UD (including like-for-not-like High Priority species that are found
in the offsets but not on the mine site). For individual High Priority species,
area-based predictions indicate that 59/90 will experience net gains by 2065;
for the remaining 31 (24 plants and seven animals: four vertebrates and three
invertebrates), it is intended that NPI will be met by a combination of enrichment
(e.g. by restoring species towards presumed natural densities from their currently
depleted densities; this may be more feasible for plants than for animals; trials
The project is predicted
to have a Net Positive
Impact on littoral
forest in 2065, and
a significant Net
Positive Impact on
regional forest types.
44 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
are needed) and ex situ conservation measures such as propagation in botanical
gardens. Based on area-based predictions, no plant or invertebrate is predicted to
have a residual decline greater than 18% and no vertebrate species is predicted
to have a residual decline greater than 5% of its global distribution (i.e. 5 UD)
requiring compensation through enrichment and insurance through ex situ
conservation measures.
These calculations are based on a number of assumptions; they are based on the
best data available at present although there is significant uncertainty around
any prediction of what will happen 55 years hence.
A key assumption that makes a significant difference to the analysis is
the counterfactual scenario considered. We propose that a precautionary
counterfactual scenario (0.9%, the Madagascar national average) is the most
appropriate to use. The analyses presented here show that, even if only a very
low rate of deforestation is presumed (Figure 7), the Rio Tinto QMM project
will have a Net Positive Impact on littoral forest over the life of the mine based
on the proposed mitigation portfolio.
5.2 Achieving NPI—what does it mean for
Rio Tinto QMM?
One of the original reasons for carrying out this analysis was to provide transparent
quantification of likely losses and gains, in order to inform debate and allow a
consensus to be reached between Rio Tinto QMM and the Biodiversity Committee
regarding what ‘achieving NPI’ means for Rio Tinto QMM.
Following discussion of an earlier draft version of this report at the Biodiversity
Committee Meeting held on 3-6 May 2010 in Fort Dauphin, Madagascar, the
Biodiversity Committee recommended that ‘achieving NPI’ means achieving a
Net Positive Impact on littoral forest
48
(using QH as a metric) and achieving a
Net Positive Impact on species (using UD of High Priority species as a metric).
The target is to achieve NPI for each High Priority species individually. Measures
of UD should be based on population size as well as distribution area where
appropriate; it is likely that population information will become available for
more species as time goes on. It is assumed that QH will provide an adequate
surrogate for other species. Rio Tinto QMM accepted these recommendations.
It is important to note that QH and UD are the main metrics used to forecast
NPI over the mine life, and to plan for the type and scale of interventions
required, but they are not the only metric or indicator that will be used at Rio
Tinto QMM. In addition, to track progress against broader conservation goals
48 I.e. offsets in a different forest type, such as the humid forest offset at Tsitongambarika, can
act as an ‘insurance policy’ and a kind of additional benefit (as well as providing key ecosystem
services such as water for local communities), but it was agreed that a loss of littoral forest cannot
be compensated for by a gain in a different habitat type. This decision was made at QMM based
on the local context; however at other sites a different decision might be made.
One of the original
reasons for carrying
out this analysis was to
provide transparent
quantification of likely
losses and gains.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 45
as outlined in the Biodiversity Action Plan (Rio Tinto QMM, 2010), Rio Tinto
QMM and partners will be monitoring a number of indicators of e.g. ecosystem
integrity, diversity of common species, forest regeneration, presence/absence and
numbers of invasive species, etc. Further details can be found in the Biodiversity
Action Plan and in specific monitoring protocols developed (or currently under
development) in consultation with the Biodiversity Committee.
5.3 Rio Tinto QMM’s biodiversity offsets
Figure 8 shows sites that are current Rio Tinto QMM biodiversity offsets, or
that are under consideration as Rio Tinto QMM offsets. Three offset sites were
included in the quantitative analysis in this report; two of these sites (Ste Luce
Forests and Mahabo) already have active conservation programmes, while the
third (Bemangidy) is awaiting confirmation as an offset by Rio Tinto QMM and
the Biodiversity Committee.
Two further potential offset sites are discussed below (Ambatotsirongorongo and
TGK I Direct Payments Project)both of these sites have had active conservation
projects funded by Rio Tinto QMM over several years, but review is ongoing
regarding whether they should be maintained as offset sites in the future.
Consequently they have not been included in the quantitative NPI forecast in
this report.
Ste Luce Forests are 500 ha of littoral forest, adjacent to (but not overlapping)
the Ste Luce ilmenite deposit. They are very similar to forests on the Rio Tinto
QMM deposits, and so are a like-for-like offset for both habitat type and species.
They are part of the mining leases so Rio Tinto QMM has greater control than
at other offset sites (reducing uncertainty). A conservation programme has
been operating at this site since before 2004, initiated, funded and managed
by Rio Tinto QMM.
Mahabo is a 1,500 ha littoral forest offset site located several hundred kilometres
north of Fort Dauphin. In collaboration with Missouri Botanical Gardens it
was chosen as a ‘like-for-like’ offset site for littoral forest as a habitat type. Its
distant location is not ideal but it represents the best option, after Ste Luce, for
a littoral forest habitat offset through averted deforestation and degradation.
A conservation and development programme with local communities has been
operating since 2004 at this site (jointly funded and managed by Rio Tinto QMM
and Missouri Botanical Gardens).
Bemangidy (Tsitongambarika III) is a c.4,000 ha parcel of lowland humid
forest within the larger c.60,000 ha Tsitongambarika forest. It is a like-for-not-like
offset for locally endemic species, and a like-for-like offset for more widespread
species. Additionally, it provides important ecosystem service benefits such as
water provision and carbon sequestration. Bemangidy currently has the status
of a proposed offset sitea Management Plan has been drafted for the site.
As a large parcel of remaining forest, even though it is not a ‘like-for–like’
Bemangidy provides
important ecosystem
service benefits such as
water provision and
carbon sequestration.
46 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
offset, it provides a crucial ‘insurance policy’ in case conservation at some of
the smaller sites is less successful than hoped in the long term. Even with active
conservation intervention, small parcels of forest can be difficult to protect and
manage in the long term.
In 2004, through discussions with BirdLife International and Asity Madagascar
(the BirdLife partner in Madagascar), the value of the whole Tsitongambarika
unprotected area of humid forest was identified as a strategic priority for Rio
Tinto QMM. It was clear the area had the potential to act as an offset site for
many littoral forest species and also had huge value as the regional ecosystem
services hub for water, soil fertility, non-timber forest products, littoral forest
pollination and seed dispersal, and local climate regulation. Significant conservation
activities have already been implemented with the involvement and support of
Rio Tinto QMM that benefit the whole forest.
1. Policy and legislation: protected area designation under national law. Rio
Tinto QMM and Asity Madagascar have carried out the preparatory work
necessary to have Tsitongambarika designated a new national protected
area, within the Système des Aires Protegées de Madagascar. This work
was funded by Rio Tinto QMM. Tsitongambarika currently has temporary
protection status N°21480 of 02/12/2008, pending political processes.
2. Long-term planning: a Tsitongambarika Conservation Management
Plan. A coalition of agents including government, CBOs, NGOs and local
communities has collaboratively developed a long-term management plan
for the forest.
3. A community-based forest conservation project has been implemented in
Tsitongambarika 1 by Asity Madagascar (funded by Rio Tinto QMM); this is
described further under ‘Tsitongambarika 1’ below.
4. Ongoing surveys of flora and fauna since 2003.
Rio Tinto QMM is currently deciding where in Tsitongambarika its offset activities
should be focused. At present, the most likely strategy is that Rio Tinto QMM’s
offset activities will be focused in a particular area within Tsitongambarika (e.g.
Bemangidy or similar), but that Rio Tinto QMM will seek partner organizations
(such as Asity Madagascar, which is the lead responsible party for the whole
protected area under government policy) and co-financing to carry out conservation
activities throughout the whole of Tsitongambarika.
Tsitongambarika 1 is the southern part of the large forest massif that also
encompasses Bemangidy. A Direct Payments community-based avoided
deforestation programme was operated by Asity Madagascar in six villages in
this area from 2007 to 2009, funded by Rio Tinto and QMM. It is an incentive-
based community conservation project whereby forest integrity is conserved in
exchange for local community development such as schools and clinics. This was
a pilot project and longer-term involvement of Rio Tinto QMM in Tsitongambarika
as a whole is currently under review, as described above.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 47
Ambatotsirongorongo is an area of transitional forest to the west of Petriky
which has suffered a high rate of deforestation. The forest is home to a significant
population of Phelsuma antanosy (Critically Endangered) currently only otherwise
found in the littoral forests in Ste Luce (including the Ste Luce Forests offset).
This small gecko was previously also observed in Petriky, but background habitat
loss through subsistence agriculture appears to have eliminated this population.
The Ambatotsirongorongo programme is a partnership between Rio Tinto QMM,
the Wildlife Conservation Society and local communities.
5.4 Monitoring forest loss
A significant proportion of biodiversity gains for Rio Tinto QMM will accrue
through averted loss. For Rio Tinto QMM to be able to credibly claim these gains
in 2065, a robust and consistent monitoring system will be needed to document
rates of forest loss across the mining leases and all the offset sites, to show that
real measurable averted loss is occurring. It would also be advisable to measure
rates of loss in the wider region over the same period to provide context and to
help monitor Rio Tinto QMM’s success in tackling potential secondary impacts
such as habitat conversion resulting from in-migration. Measuring averted
deforestation will be a key piece of NPI accounting in the future, and needs to be
given careful thought at an early stage in project planning. Whatever method is
chosen, care must be taken that it does not provide a ‘perverse incentive’ for Rio
Tinto QMM to encourage (or fail to help reduce) forest loss in the wider region.
Measuring averted
deforestation will
be a key piece of
NPI accounting in
the future, and
needs to be given
careful thought at
an early stage in
project planning.
48 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Figure 9.
Current biodiversity offset sites or sites that are under consideration
as offsets.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 49
5.5 The broader context—drivers of biodiversity
loss in southeastern Madagascar
As in other parts of the world, forest conservation in southeastern Madagascar
will only be successful in the long term if the underlying drivers of biodiversity
loss are addressed. Prior to the arrival of Rio Tinto QMM in the 1990s, the littoral
forest was already in rapid decline (Vincelette et al., 2007b). Apart from mining,
other major causes of current forest loss are unsustainable levels of charcoal
production and tavy (slash-and-burn agriculture). People using forests in these
ways do so because they have no other option82% of Anosy inhabitants live
below the poverty line (US1$/day; Vincelette et al., 2007a).
Many of Rio Tinto QMM’s Additional Conservation Actions seek to address these
underlying causes of biodiversity loss, for example by setting up plantations of
fast-growing non-native trees to reduce charcoal production pressure on native
forest, or providing alternative livelihoods by training local people in skills such
as blacksmithing, crop production and diversification, vetiver grass production
and planting, improved rice production, animal breeding, handicrafts, etc.
However, southeastern Madagascar’s social and environmental problems cannot
be solved by the actions of a single company; broader initiatives are needed
involving multiple agencies and stakeholders.
5.6 NPI analysis updates and adaptive management
As part of Rio Tinto QMM’s biodiversity action planning process, this analysis
and forecast should be periodically revisited to take into account changes such
as newly discovered or described species, new listings (and up- or down-listings)
on the IUCN Red List, and methodological improvements that can be made as
the state of knowledge and best practice in relevant fields such as biodiversity
metrics and biodiversity offsetting evolve. This will help Rio Tinto QMM to track
progress and will contribute to adaptive management. It will also allow the
assumptions made in the present analysis to be tested based on empirical data
(e.g. is it reasonable to assume that deforestation can be completely stopped
in the Avoidance Zones? Is it reasonable to assume that conservation actions
will improve the quality score of the Avoidance Zone in Petriky by 0.05 every 15
years?). It is recommended that this analysis should be revisited at least every
five years (the same periodicity is recommended for revisiting the Biodiversity
Action Plan: Rio Tinto QMM, 2010). Section 7 details lessons learned in carrying
out the present analysis and recommends areas for future researchsubsequent
iterations of this NPI forecast should take into account these lessons, as well as
any relevant new research findings.
Forest conservation
in southeastern
Madagascar will only
be successful in the
long term if the
underlying drivers of
biodiversity loss are
addressed.
50 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
6 LESSONS LEARNED
AND DIRECTIONS FOR
FUTURE RESEARCH
6.1 Potential impacts of climate change
An important point to consider in the future is how Rio Tinto QMM can factor
the potential impacts of climate change into biodiversity action planning and
implementation of restoration and offsets. The impact sites, avoidance and
restoration zones, and littoral forest offset sites are all close to the coast and
at very low altitudes. Consequently, by 2065, they may plausibly be impacted
by sea level rise (especially when associated with storm surges), and will face
increased temperatures and decreased rainfall. This could pose problems for the
long-term viability of Rio Tinto QMM’s littoral forest conservation areas. Among
many other things it may affect the ease and effectiveness of restoration, and
could wipe out other gains in Quality Hectares as the climate becomes less
suitable for littoral forest or the coast is inundated. There may be a need to
secure and re-vegetate areas that are further inland or upslope, and to think
about buffers and corridors for existing sites. This underlines the importance of
Rio Tinto QMM’s ongoing monitoring of the habitats and species and adaptive
management to changing conditions as they occur (Rio Tinto QMM, 2010).
Recent research modelling species range shifts in Madagascar in response to
future climate change predicted that the littoral forest will disappear (Hannah
et al., 2008), and palaeoecological reconstruction indicates that littoral forest
in the Fort Dauphin region has been highly dynamic in response to climatic
changes in the past (Virah-Sawmy et al., 2010). Given that the Rio Tinto QMM
mine is planning for the 2004–2065 time period, it would be useful to evaluate
the vulnerability of the conservation actions described in this report to different
climate change scenarios (i.e. less rain, warmer temperatures, higher sea level,
increased frequency of storms) to determine how resilient the proposed activities
would be. It would also be informative to supplement the broad-scale modelling
of Hannah et al. (2008) with a more detailed study at the regional level.
6.2 Defining a standard set of species to be
included in NPI accounting
At present, Rio Tinto’s guidance on which species to include in NPI accounting is
not prescriptive, although it does state that this decision should be based on the
classic conservation priority-setting principles of vulnerability and irreplaceability,
and should be done in consultation with relevant experts and stakeholders
(Ekstrom and Anstee, 2007; Rio Tinto, 2010). For this analysis, and following
these principles, Rio Tinto QMM (in consultation with the biodiversity committee
Recent research
modelling species
range shifts in
Madagascar in
response to future
climate change
predicted that the
littoral forest will
disappear.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 51
and other experts) determined that High Priority species as defined in the
Biodiversity Action Plan (Rio Tinto QMM, 2010) should be included. These
include globally Critically Endangered and Endangered species and species with
a highly restricted distribution (see Rio Tinto QMM, 2010 for a definition), but
excluded Vulnerable species.
One of the lessons learned from this process is that it may be valuable for Rio
Tinto to define a standard set of species to be included in NPI accounting, based
on criteria that are applied in the same way to all Rio Tinto sites globally. This
should likely include Vulnerable species as well as those listed as Endangered and
Critically Endangered on the IUCN Red List. It may also be valuable to establish
standard definitions for what constitutes ‘restricted range’. If Rio Tinto develops
such standard guidelines, they should be used in subsequent iterations of the
present analysis.
6.2.1 Undescribed species
When it comes to defining a standard set of species for inclusion in NPI accounting,
undescribed species present an interesting conundrum. The present analysis
includes a number of possible new species that have yet to be formally described.
Many of these will be described over the course of the project, and Rio Tinto
QMM arguably has a responsibility to support the taxonomic research to allow
this. However, it is debatable whether it is appropriate to include such taxa in
formal NPI calculations. Based on experience from carrying out this analysis, we
would recommend that Rio Tinto develops a standard policy for dealing with
undescribed species.
6.3 Counterfactual scenarios, baselines, and
calculating biodiversity gains from averted loss
This is a particularly fertile and fast-moving field of research at the moment
given interest in REDD (Reducing Emissions from Deforestation and Degradation)
and growing interest in biodiversity offsets. It will be important to keep abreast
of developments in this field, and evaluate their relevance to calculation of
biodiversity offset gains at the site level, to ensure that NPI accounting and
forecasting is as robust as possible. Specific questions raised elsewhere in this
report include how to ensure that calculation methods do not inadvertently
create perverse incentives, and how best to monitor deforestation and measure
gains going forward.
6.4 Calculating Units of Global Distribution—
methods for measuring species distribution area
In the present analysis, in cases where UD was measured based on distribution
rather than population (e.g. where population data were unavailable at the global
and site level), it was necessary to estimate the global distribution area for each
species. Different methods were used depending on the taxon in question. For
When it comes to
defining a standard set
of species for inclusion
in NPI accounting,
undescribed species
present an interesting
conundrum.
52 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
example, for some mammals and birds area of occupancy (AOO) was calculated
by measuring extent of occurrence (EOO) based on maps from the IUCN Red
List of Threatened Species (IUCN, 2009), and inferring AOO based on a very
rough ‘rule of thumb’ of 10% of EOO (this relationship between AOO and
EOO is implied by the thresholds set for Criterion B of the IUCN Red List). EOO
cannot be used without some modification because it typically overestimates
(sometimes vastly) the area occupied by each species (Jetz et al., 2008; Rodrigues,
2011). However, for future analyses of this type, it could be worth exploring
whether other metrics such as Extent of Suitable Habitat (ESH) offer a useful
alternative. ESH is the area of potentially suitable vegetation types within the
altitudinal preferences of the species (Rondinini, et al., 2005; Buchanan et al.,
2008). In a recent study using the ESH method, species’ range was restricted on
average to 28% of its original EOO (Beresford et al., 2011). However, the ESH
method is not without its problems, as detailed by Rodrigues et al. (2011); a
recent ground-truthing exercise suggested that the ESH method had very little
capacity to distinguish between occupied and non-occupied parts of species’
EOOs (Beresford et al., 2011; Rodrigues et al., 2011).
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 53
7 BIBLIOGRAPHY
BBOP Multipliers Consultation Working Group. 2008. Business and Biodiversity
Offsets Program (BBOP) Consultation paper: The use of multipliers to deal
with risk, uncertainty, time discounting and landscape conservation targets in
biodiversity offset design. Unpublished report.
Business and Biodiversity Offsets Programme (BBOP). 2009. Biodiversity Offset
Cost-Benefit Handbook. BBOP, Washington, DC. www.forest-trends.org/
biodiversityoffsetprogram/guidelines/cbh.pdf.
Bekessy, S.A., Wintle, B.A., Lindenmayer, D.B., Mccarthy, M.A., Colyvan, M.,
Burgman, M.A. and Possingham, H.P. 2010. The biodiversity bank cannot be a
lending bank. Conservation Letters, 3: 151–158.
Beresford, A.E., Buchanan, G.M., Donald, P.F., Butchart, S.H.M., Fishpool, L.D.C.
and Rondinini, C. 2011. Poor overlap between the distribution of protected
areas and globally threatened birds in Africa. Animal Conservation, 14: 99–107.
Buchanan, G.M., Butchart, S.H.M., Dutson, G., Pilgrim, J.D., Steininger, M.K.,
Bishop, K.D. and Mayaux, P. 2008. Using remote sensing to inform conservation
status assessment: estimates of recent deforestation rates on New Britain and
the impacts upon endemic birds. Biol. Conserv., 141: 56–66.
Burgin, S. 2008. BioBanking: an environmental scientist’s view of the role of
biodiversity banking offsets in conservation. Biodiversity and Conservation,
17: 807–816.
Consiglio, T., Schatz G.E., McPherson G., Lowry, P.P., Rabenantoandro, J., Rogers,
Z.S., Rabevohitra, R. and Rabehevitra, R. 2006. Deforestation and plant diversity
of Madagascar’s littoral forests. Conservation Biology, 20: 1799–1803.
Comité Régional pour le Développement de l’Anosy (CRD). 2003. Schéma
de Développement Régional (SRD) de l’Anosy. Rapports Phases 1 à 4. Dobbin
International Inc. Washington, DC.
Du Puy, D. and Moat, J. 1998. Vegetation Mapping and Classification in
Madagascar (Using GIS): Implications and Recommendations for the Conservation
of Biodiversity. Pp. 97-–117 in: Huxley, R., Lock, J.M. and Cutler, D.F. (eds.).
Chorology, Taxonomy and Ecology of the Floras of Africa and Madagascar. Royal
Botanic Gardens Kew, Richmond.
Dumetz, N. 1999. High plant diversity of lowland rainforest vestiges in eastern
Madagascar. Biodiversity and Conservation, 8: 273–315.
54 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Ekstrom, J.M.M. and Anstee, S. 2007. The Rio Tinto Biodiversity Values Identification
and Prioritisation System. (Unpublished internal Rio Tinto guidance note).
Ganzhorn, J.B. 1998. Nested patterns of species composition and its implications
for lemur biogeography in Madagascar. Folia Primatologica, 69: 332–341.
Ganzhorn, J.U., Lowry II, P.P., Schatz. G.E. and Sommer, S. 2001. The biodiversity
of Madagascar: one of the world’s hottest hotspots on its way out. Oryx,
35: 346–348.
Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). 2007. Biodiversity,
ecology and conservation of littoral ecosystems in southeastern Madagascar,
Tolagnaro (Fort Dauphin). Smithsonian Institution, Washington, DC.
Goodman, S.M., Pigeon, M., Hawkins, A.F.A. and Schulenberg, T.S. 1997. The
birds of southeastern Madagascar. Fieldiana: Zoology, new series, 87: 1–132.
Goodman, S.M. and Ramanamanjato J.-B. 2007. A perspective on the paleo-
ecology and biogeography of extreme southeastern Madagascar, with special
reference to animals. Pp. 25–48 in: Ganzhorn, J.U., Goodman, S.M. and
Vincelette, M. (eds.). Biodiversity, ecology and conservation of littoral ecosystems
in southeastern Madagascar, Tolagnaro (Fort Dauphin). Smithsonian Institution,
Washington, DC.
de Gouvenain, R.C. and A.J. Silander Jr. 2000. Diversité et Endémisme a
Madagascar. Pp. 45–62 in: Lourenço, W.R. and Goodman, S.M. (eds.). Mémoires
de la Société de Biogéographie. Paris.
Hannah, L., Dave R., Lowry P.P. et al. 2008. Climate change adaptation for
conservation in Madagascar. Biology Letters, 4: 590–594.
Harper, G.J., Steininger, M.K., Tucker, C.J., Juhn, D. and Hawkins, F. 2007. Fifty
years of deforestation and forest fragmentation in Madagascar. Environmental
Conservation, 34: 325–333.
Henderson, S. 1999. Relationship between structure and composition in the
littoral forests of southeast Madagascar. PhD thesis, Environmental Change
Unit, Oxford University.
International Monetary Fund (IMF). 2011. World Economic Outlook Database–April
2011, International Monetary Fund [Data refer to the year 2010]. Accessed on
11 April 2011. http://imf.org/external/pubs/ft/weo/2011/01/weodata/index.aspx.
Ingram, J.C. and Dawson, T.P. 2005. Climate change impacts and vegetation
response on the island of Madagascar. Philosophical Transactions of the Royal
Society, London, A, 363: 55–59.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 55
Ingram, J.C., Dawson, T.P. and Whittaker, R.J. 2005a. Mapping tropical forest
structure in southeastern Madagascar using remote sensing and artificial neural
networks. Remote Sensing of Environment, 94: 491–507.
Ingram J.C., Whittaker, R.J. and Dawson, T.P. 2005b. Tree structure and diversity
in human-impacted littoral forests, Madagascar. Environmental Management,
35: 779–798.
Ingram, J.C. and Dawson, T.P. 2006. Forest cover, condition, and ecology in
human-impacted forests, south-eastern Madagascar. Conservation and Society,
4: 194–230.
IUCN. 2009. 2009 IUCN Red List of Threatened Species. International Union for
Conservation of Nature, Gland, Switzerland and Cambridge, UK.
Jacques Whitford Inc. 2007. Marine Environmental Monitoring of Ehoala Port.
Unpublished report for Rio Tinto QMM.
Jetz, W., Sekercioglu, C.H. and Watson, J.E.M. 2008. Ecological correlates
and conservation implications of overestimating species geographic ranges.
Conservation Biology, 22: 110–119.
Langhammer, P.F., Bakarr, M.I., Bennun, L.A., Brooks, T.M., Clay, R.P., Darwall,
W., De Silva, N., Edgar, G.J., Eken, G., Fishpool, L.D.C., Fonseca, G.A.B. da,
Foster, M.N., Knox, D.H., Matiku, P., Radford, E.A., Rodrigues, A.S.L., Salaman,
P., Sechrest, W. and Tordoff, A.W. 2007. Identification and Gap Analysis of Key
Biodiversity Areas: Targets for Comprehensive Protected Area Systems. IUCN,
Gland, Switzerland.
Lowry II, P.P. and Faber-Langendoen, D. 1991. Overall flora and vegetation study.
In: Ilmenite Project. Social and Environmental Impact Assessment. Supporting
Document no. 4. Unpublished document presented by QIT Madagascar Minerals
S.A. (QMM S.A.) to the Madagascar National Environment Office in May 2001.
Lowry II, P.P., Smith, P.P. and Rabevohitra, R. 2001. Review of MIR Télédétection
Inc.: Deforestation Study in the region of Fort Dauphin (Tolagnaro) Madagascar.
In: Ilmenite Project. Social and Environmental Impact Assessment. Supporting
Document no. 9. Unpublished document presented by QIT Madagascar Minerals
S.A. to the Madagascar National Environment Office in May 2001.
McKenney, B. and Kiesecker J.M. 2010. Policy Development for Biodiversity Offsets:
A Review of Offset Frameworks. Environmental Management, 45: 165–176
Mittermeier, R.A., Robles Gil, P., Hoffmann, M., Pilgrim, J., Brooks, T., Mittermeier,
C.G., Lamoreux, J. and da Fonseca, G.A.B. 2004. Hotspots Revisited. CEMEX,
Mexico City.
56 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Nussbaum, R.A., Raxworthy, C.J., Raselimanana, A.P. and Ramanamanjato, J.-B.
1999. Amphibians and reptiles of the Réserve Naturelle Intégrale d’Andohahela,
Madagascar. In: Goodman, S.M. (ed.). A floral and faunal inventory of the Réserve
Naturelle Intégrale d’Andohahela, Madagascar: With reference to elevational
variation. Fieldiana: Zoology, 94: 155–173.
Parkes, D., Newell, G. and Cheal, D. 2003. Assessing the quality of native
vegetation: the ‘habitat hectares’ approach. Ecological Management and
Restoration, 4: S29–S38.
Rabenantoandro, J. 2001. Floral, structural and dendrometric analysis of low
altitude humid forest on the western slopes of the Masaola National Park
[Analyse floristique, structurale et dendrométrique de la forêt dense humide
de basse altitude du Parc National Masoala sur le versant occidental selon la
méthode de la parcelle permanente.] Dissertation study, Faculty of Sciences,
Antananarivo University.
Rabenantoandro, J., Randriatafika, F. and Lowry II, P.P. 2007. Floristic and Structural
Characteristics of Remnant Littoral Forest Sites in the Tolagnaro Area. Pp. 65-
93 in: Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). Biodiversity,
Ecology and Conservation of Littoral Ecosystems in Southeastern Madagascar;
Tolagnaro (Fort Dauphin). Smithsonian Institution, Washington DC.
Rabevohitra, R., Wilmé, L., Lowry II, P.P. and Schatz, G.E. 1998. Pp. 65–99 in:
Ratsirarson, J. and Goodman, S.M. (eds.). Biological inventory of the littoral forest
of Tampolo. Volume 14. [Inventaire biologique la Foret Littorale de Tampolo.]
Volume 14. Centre d’Information et de Documentation Scientifique et Technique,
Antananarivo, Madagascar.
Ramsar Convention Secretariat. 2004. Designating Ramsar Sites. Strategic
Framework and Guidelines for the Future Development of the List of Wetlands
of International Importance. Second edition. Ramsar Convention Secretariat,
Gland, Switzerland.
Ramanamanjato, J.B., Mcintyre, P.B. and Nussbaum, R.A. 2002. Reptile, amphibian
and lemur diversity of the Malahelo Forest, a biogeographical transition zone
in southeastern Madagascar. Biodiversity and Conservation, 11: 1791–1807.
Rarivoson, C. and Mara, R. 2007. Testing of Fast Growing Tree Species for the
Rehabilitation of Mine Sites. Pp. 323–331 in: Ganzhorn, J.U., Goodman, S.M.
and Vincelette, M. (eds.). Biodiversity, Ecology and Conservation of Littoral
Ecosystems in Southeastern Madagascar; Tolagnaro (Fort Dauphin). Smithsonian
Institution, Washington DC.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 57
Raxworthy, C.J. 2003. Introduction to the Reptiles. Pp. 934–949 in: Goodman,
S.M. and Benstead, J.P. (eds.). The Natural History of Madagascar. (Chicago
Press, Chicago.)
Rio Tinto. 2004. Rio Tinto’s biodiversity strategy: Sustaining a natural balance.
Rio Tinto, London and Melbourne. Unpublished document available at: www.
riotinto.com/documents/ReportsPublications/RTBidoversitystrategyfinal.pdf.
Rio Tinto. 2008a. Rio Tinto and biodiversity: Achieving results on the ground.
Rio Tinto, London and Melbourne. Unpublished document available at: www.
riotinto.com/documents/ReportsPublications/RTBidoversitystrategyfinal.pdf.
Rio Tinto. 2008b. Standard E9Land Use Stewardship. Unpublished
document available at: www.riotinto.com/documents/ReportsPublications/
LandUseStewardship.pdf.
Rio Tinto. 2008c. A biodiversity case study of Rio Tinto Madagascar / QMM:
Achieving a Net Positive Impact on Biodiversity.
Rio Tinto. 2009. The Way We Work: our global code of business conduct.
Unpublished document available at: www.riotinto.com/documents/The_way_
we_work.pdf.
Rio Tinto. 2010. Biodiversity Action Planning Guidance Note. Unpublished document.
Rio Tinto QMM. 2001. Social and Environmental Impact Assessment “Ilmenite
Project” (Office National d’Environnement 2001).
Rio Tinto QMM. 2008. Biodiversity Offsets StrategyQMM Rio Tinto Madagascar:
Draft working paper and issues for discussion. Unpublished document.
Rio Tinto QMM. 2010. Rio Tinto QMM Biodiversity Action Plan.
Unpublished document.
Rodrigues, A.S.L. 2011. Improving coarse species distribution data for conservation
planning in biodiversity-rich, data-poor, regions: no easy shortcuts. Animal
Conservation, 14: 108–110.
Rondinini, C., Stuart, S. and Boitani, L. 2005. Habitat suitability models and
the shortfall in conservation planning for African vertebrates. Conservation
Biology, 19: 1488–1497.
Schütte, K. 2007. An inventory of the Mantodea and Phasmatodea in the
littoral forests near Tolagnaro (Fort-Dauphin) / Inventaire des Mantodea et des
Phasmatodea dans les forêts littorales des environs de Tolagnaro (Fort-Dauphin).
Pp. 155–162 in: Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.).
58 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Biodiversity, Ecology and Conservation of Littoral Ecosystems in Southeastern
Madagascar; Tolagnaro (Fort Dauphin). Smithsonian Institution, Washington DC.
Schütte, K. and Razafindraibe, P. 2007. Checklist of dragonflies of the littoral
forests near Tolagnaro (Fort-Dauphin) / Liste de contrôle des libellules des
forêts littorales des environs de Tolagnaro (Fort-Dauphin). Pp. 163–166 in:
Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). Biodiversity, Ecology
and Conservation of Littoral Ecosystems in Southeastern Madagascar; Tolagnaro
(Fort Dauphin). Smithsonian Institution, Washington DC.
Society for Ecological Restoration International Science & Policy Working Group.
2004. The SER International Primer on Ecological Restoration. Society for Ecological
Restoration International, Tucson, USA.
TEEB. 2010. The Economics of Ecosystems and Biodiversity: Mainstreaming the
Economics of Nature: Chapter 4 Scaling down biodiversity and ecosystem risks
to business www.teebweb.org/LinkClick.aspx?fileticket=tcneop1kys4%3d&ta
bid=1021&language=en-US.
ten Kate, K., Bishop, J. and Bayon, R. 2004. Biodiversity offsets: Views, experience,
and the business case. IUCN, Gland, Switzerland and Cambridge, UK and Insight
Investment, London.
van Aarde, R.J., Ferreira, S.M., Kritzinger, J.J., van Dyk, P.J., Vogt, M. et al. 1996.
An evaluation of habitat rehabilitation on coastal dune forests in northern
KwaZulu-Natal, South Africa. Restoration Ecology, 4: 334–345.
Vincelette, M., Dean, L. and Ganzhorn, J.U. 2007a. The QMM/Rio Tinto project
history in Tolagnaro and its social and environmental concepts. Pp. 1–8 in:
Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). Biodiversity, Ecology
and Conservation of Littoral Ecosystems in Southeastern Madagascar; Tolagnaro
(Fort Dauphin). Smithsonian Institution, Washington DC.
Vincelette, M. Theberge, M. and Randrihasipara, L. 2007b. Evaluations of forest
cover at regional and local levels in the Tolagnaro region since 1950. Pp. 49–59
in: Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). Biodiversity, Ecology
and Conservation of Littoral Ecosystems in Southeastern Madagascar; Tolagnaro
(Fort Dauphin). Smithsonian Institution, Washington DC.
Vincelette, M., Dumouchel, J., Giroux, J. and Heriarivo, R. 2007c. The Tolagnaro
(Fort Dauphin) region: a brief overview of the geology, hydrology and climatology.
Pp. 9–18 in: Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.). Biodiversity,
Ecology and Conservation of Littoral Ecosystems in Southeastern Madagascar;
Tolagnaro (Fort Dauphin). Smithsonian Institution, Washington DC.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 59
Vincelette, M. Rabenantoandro, J., Randrihasipara, L., Randriatafika, F. and
Ganzhorn, J. 2007d. Results from Ten Years of Restoration Experiments in the
Southeastern Littoral Forests of Madagascar. Pp. 337–354 in: Ganzhorn, J.U.,
Goodman, S.M. and Vincelette, M. (eds.). Biodiversity, Ecology and Conservation
of Littoral Ecosystems in Southeastern Madagascar; Tolagnaro (Fort Dauphin).
Smithsonian Institution, Washington DC.
Virah-Sawmy, M. and Ebeling, J. 2010. The difficult road toward real-world
engagement: conservation science and mining in southern Madagascar.
Conservation Letters, 3: 288–289.
Virah-Sawmy, M., Willis, K.J. and Gillson, L. 2010. Evidence for drought and
forest declines during the recent megafaunal extinctions in Madagascar. Journal
of Biogeography, 37(3): 506–519.
Watson, J.E.M., Whittaker, R.J. and Dawson, T.P. 2005. The importance of littoral
forest remnants for indigenous bird conservation in southeastern Madagascar.
Biodiversity and Conservation, 14: 523–545.
Watson, J.E.M. 2007. Conservation of Bird Diversity in Madagascar’s Southeastern
Littoral Forests. Pp. 186–208 in: Ganzhorn, J.U., Goodman, S.M. and Vincelette,
M. (eds.). Biodiversity, Ecology and Conservation of Littoral Ecosystems in
Southeastern Madagascar; Tolagnaro (Fort Dauphin). Smithsonian Institution,
Washington DC.
Wesener, T. 2009. Unexplored richness: discovery of 31 new species of Giant Pill-
Millipedes endemic to Madagascar, with a special emphasis on microendemism
(Diplopoda, Sphaerotheriida). Zootaxa, 2097: 3–134.
Wesener, T. and Wägele, J.W. 2007. Giant Pill-Millipedes (Diplopoda:
Sphaerotheriida) in the Littoral Forest Fragments of Southeastern Madagascar.
Pp. 147–154 in: Ganzhorn, J.U., Goodman, S.M. and Vincelette, M. (eds.).
Biodiversity, Ecology and Conservation of Littoral Ecosystems in Southeastern
Madagascar; Tolagnaro (Fort Dauphin). Smithsonian Institution, Washington DC.
Wesener, T., Enghoff, H. and Sierwald, P. 2009. Review of the Spirobolida on
Madagascar, with descriptions of twelve new genera, including three genera
of ‘fire millipedes’ (Diplopoda). Zookeys, 19: Special Issue.
60 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 61
APPENDIX 1
QMM Biodiversity Committee
QMM invited external biodiversity experts to participate in a formal Biodiversity
Committee in 2003. The Committee had operated informally for several years
prior to this. The Committee operates with full autonomy, being free to publically
criticize Rio Tinto QMM (either individually or collectively), and receiving no
remuneration for time spent on Committee meetings.49
The purpose of the Committee is to advise QMM on how best to conserve and
enhance biodiversity within the project area before, during and after mining.
The Committee also advises QMM and the regional authorities on biodiversity
issues within the Fort Dauphin area. The Committee assists QMM with the
implementation of its Biodiversity Programme, for example by critically reviewing
and providing input to the draft Biodiversity Action Plan (Rio Tinto QMM, 2010)
and the present report, and assisting with the preparation of the biodiversity
monograph (Ganzhorn et al., 2007).
The Guiding Principles of operation for the Biodiversity Committee include:
• Open and honest communication;
• Non-attribution of discussions;
• Full access to social and environmental information;
• Information and Committee process accessible to public;
• Members retain freedom to communicate, singly or collectively, about
Committee activities and the project.
The Committee meets once a year; additional ad hoc meetings may be arranged
between some or all members at other times. Meetings normally last three
days, and include a mixture of technical and strategic issues. The Committee
provides expert advice to QMM and can identify its own topics of interest and
agenda priorities. The last day of each meeting includes a ‘closed session’ that
only the Committee can attend (e.g. no Rio Tinto QMM staff may be present).
The Committee typically gives Rio Tinto QMM a set of formal comments and
recommendations at the end of each meeting, to which QMM must respond.
QMM provides for any additional work requested by the Committee and
mutually agreed upon.
Committee meetings are hosted by QMM; travel and living expenses related
to meetings are reimbursed by QMM; however Committee members are not
remunerated for their time and contribution to Committee meetings.
49 Travel and living expenses are reimbursed, and committee members (or their institutions) may
at times be engaged to deliver specific pieces of work which are financed either as separate
contracts, or through the existing partnership frameworks (e.g. Rio Tinto has formal partnership
programmes w ith a number of NGOs including RBG Kew), as appropriate.
62 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
APPENDIX 2
Quality Hectares and Units of
Global Distribution
Quality Hectares (QH)
Quality Hectares are Rio Tinto’s standard metric for tracking progress towards
the NPI target at the global and site level. They are conceptually related to
the ‘habitat hectares’ metric used in Victoria, Australia (Parkes et al., 2003). A
wide range of biodiversity values, including threatened species, rare habitats
or non-timber forest products, may be expressed in terms of their quantity and
quality. This is expressed as an ‘Area × Quality’ metric, referred to here as Quality
Hectares (QH). For example, 100 hectares of forest in pristine condition would
count as 100 Quality Hectares (100 ha × 100% quality = 100 QH), whereas 100
hectares of fairly degraded forest at 40% ‘optimal quality’ would be expressed
as 40 Quality Hectares (100 ha × 40% quality = 40 QH).
Quality or condition can be measured in a variety of different ways, depending
upon the habitat in question. Rio Tinto does not have a single ‘one-size-fits-all’
global method for measuring habitat quality because it is so context dependent,
but rather recommends using established and accepted methods where these
are available at the regional (e.g. national or state) level. Because there was no
established methodology for evaluating habitat condition in Madagascar, QMM
has developed its own classification method, with input from Missouri Botanical
Garden (MBG), Royal Botanic Gardens (RBG Kew), FOFIFA,50 and other experts.
This method is described in Vincelette et al., (2007b) and Rabenantoandro et
al. (2007).
Units of Global Distribution (UD)
Units of Global Distribution are a novel metric, developed for this analysis, but
conceptually related to Quality Hectares. A Unit of Global Distribution is equivalent
to 1% of a species’ global population (or 1% of its global distribution, in the
event that population data are unavailable). Units of Global Distribution are
calculated as follows: if a species has a global population of 1,000 individuals,
and 10 of those are lost, that would be a loss of 1% of the global population or
1 ‘Unit of Global Distribution’ (UD). Similarly, if a species has a global distribution
of 100 ha, and 1 ha of its distribution is lost as a result of habitat loss caused
by mining, that is a loss of 1% of its global distribution or 1 ‘Unit of Global
Distribution’ (UD).
A precedent for the use of this kind of metric can be found in the quantitative
thresholds used to identify sites of global significance for biodiversity conservation.
Under the Convention on Wetlands (Ramsar Convention, 1971), sites are
50 The National Centre for Applied Research into Rural Development, w ww.fofifa.mg.
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 63
selected for the List of Wetlands of International Importance according to a suite
of criteria adopted by the Conference of Parties, one of these being Criterion
6: “A wetland should be considered internationally important if it regularly
supports 1% of the individuals in a population of one species or subspecies of
waterbird” (Ramsar Convention Secretariat, 2004). Similarly, the criteria for the
identification of Key Biodiversity Areas (KBAs) use thresholds of 1% or 5% of
the species’ global population (depending on the type of species in question)
to identify Key Biodiversity Areas (Langhammer et al., 2007).
This is the first offsets analysis to use Units of Global Distribution (UD) as a metric
or currency, so their use warrants a brief discussion. Crucially, the use of the
‘Units of Global Distribution’ metric should not be seen to imply that different
species are directly exchangeable. It will never be acceptable to simply state
that it is permissible to render a species of plant extinct so long as one protects
in perpetuity the entire global range of a particular frog species. Rather, it gives
an idea of the scale of losses, and the concomitant scale of gains required to
give a Net Positive Impact.
The advantage of Units of Global Distribution is that this metric is much more
closely linked to the extinction risk51 faced by a species than is simple hectares
(or even Quality Hectares). For example, a loss of 80 ha of habitat would be
catastrophic for a species with a total global distribution of 100 ha, but would
be of negligible significance for a species with a total global distribution of
1,000,000 ha. The disadvantage is that it is slightly more difficult and time-
consuming to calculate, although as demonstrated by the present analysis it is
feasible. Essentially, Units of Global Distribution are a way of calibrating losses/
gains measured in hectares (or Quality Hectares) so that they take into account
the total global area of distribution of a particular biodiversity value. UD can
be calculated for species, habitats, or other types of biodiversity value (e.g.
non-timber forest products).
51 Indeed, a reduction in Units of Global Distribution is essentially the s ame as what is measured by
Criterion A of the I UCN Red List of Threatened Species. Under Criterion A, if a species loses >80%
of its global population in 10 years /3 generations (which can be inferred based on reduction in
distribution), it would be clas sed as Critically Endangered. Similarly, a reduction of >50% would
render a species Endangered and a reduction of >30% would render a species Vulnerable (IUCN,
2001) .
64 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
APPENDIX 3
Losses and gains in Quality Hectares (QH) of littoral forest
Type of loss/gain Site
Forest cover in 2004
Background
deforestation rate
Estimated forest cover
remaining in 2065 given
no mine
Quality in 2004
Total QH Lost
HA forest cover AZ
or Offset
Quality in 2004
QH in 2004
Quality in 2065
Area in 2065
QH in 2065
Deforestation rate
Reduced
deforestation rate
Years of Intervention
since 2004 baseline
QH remaining after 61
years @ background rate
QH remaining after 61
years @ 50%/75%/100%
background rate
What proportion of
offset gains can be
claimed by Rio Tinto?
Net QH of Forest
Gained through
QMM investment
LOSSES Mandena Mining 330 0.009 208 0.11 23 -23.3
LOSSES Petriky Mining 674 0.009 425 0.23 98 -97.7
LOSSES Ste Luce Mining 661 0.009 417 0.74 307 -307.1
RESTORATION
GAINS
Mandena
restoration
0.35 225 79 79
RESTORATION
GAINS
Petriky
restoration
0.20 225 45 45
RESTORATION
GAINS
Ste Luce
restoration
0.25 225 56 56
AVERTED LOSS
IN AZs
Mandena
Avoidance (AZ)
230 0.57 132 0.009 061 76.1 132.1 1.00 56.0
AVERTED LOSS
IN AZs
Petriky
Avoidance (AZ)
120 0.39 46 0.009 061 26.6 46.3 1.00 19.6
AVERTED LOSS
IN AZs
Ste Luce
Avoidance (AZ)
274 0.80 218 0.009 061 125.6 218.1 1.00 92.4
INCREASED
QUALITY IN AZs
Mandena
Avoidance (AZ)
230 0.57 132 0.97 230 224 1.00 92.0
INCREASED
QUALITY IN AZs
Petriky
Avoidance (AZ)
120 0.39 46 0.59 120 70 1.00 24.0
INCREASED
QUALITY IN AZs
Ste Luce
Avoidance (AZ)
274 0.80 218 1.00 274 274 1.00 55.9
AVERTED LOSS
IN OFFSETS
Ste Luce Offset 498 0.84 420 0.009 0.0023 61 242.0 366.1 1.00 124.1
AVERTED LOSS
IN OFFSETS
Mahabo 1565 0.80 1252 0.008 0.0040 61 767.0 980.4 0.63 134.5
AVERTED LOSS
IN OFFSETS
Bemangidy 4000 0.90 3600
0.0255
0.0128 61 744.7 1645.7 1.00 901.0
Total All
forest Types
1251.4
Total Fort
Dauphin
Littoral Forest
215.9
Total Littoral
Forest
350.4
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 65
APPENDIX 3
Losses and gains in Quality Hectares (QH) of littoral forest
Type of loss/gain Site
Forest cover in 2004
Background
deforestation rate
Estimated forest cover
remaining in 2065 given
no mine
Quality in 2004
Total QH Lost
HA forest cover AZ
or Offset
Quality in 2004
QH in 2004
Quality in 2065
Area in 2065
QH in 2065
Deforestation rate
Reduced
deforestation rate
Years of Intervention
since 2004 baseline
QH remaining after 61
years @ background rate
QH remaining after 61
years @ 50%/75%/100%
background rate
What proportion of
offset gains can be
claimed by Rio Tinto?
Net QH of Forest
Gained through
QMM investment
LOSSES Mandena Mining 330 0.009 208 0.11 23 -23.3
LOSSES Petriky Mining 674 0.009 425 0.23 98 -97.7
LOSSES Ste Luce Mining 661 0.009 417 0.74 307 -307.1
RESTORATION
GAINS
Mandena
restoration
0.35 225 79 79
RESTORATION
GAINS
Petriky
restoration
0.20 225 45 45
RESTORATION
GAINS
Ste Luce
restoration
0.25 225 56 56
AVERTED LOSS
IN AZs
Mandena
Avoidance (AZ)
230 0.57 132 0.009 061 76.1 132.1 1.00 56.0
AVERTED LOSS
IN AZs
Petriky
Avoidance (AZ)
120 0.39 46 0.009 061 26.6 46.3 1.00 19.6
AVERTED LOSS
IN AZs
Ste Luce
Avoidance (AZ)
274 0.80 218 0.009 061 125.6 218.1 1.00 92.4
INCREASED
QUALITY IN AZs
Mandena
Avoidance (AZ)
230 0.57 132 0.97 230 224 1.00 92.0
INCREASED
QUALITY IN AZs
Petriky
Avoidance (AZ)
120 0.39 46 0.59 120 70 1.00 24.0
INCREASED
QUALITY IN AZs
Ste Luce
Avoidance (AZ)
274 0.80 218 1.00 274 274 1.00 55.9
AVERTED LOSS
IN OFFSETS
Ste Luce Offset 498 0.84 420 0.009 0.0023 61 242.0 366.1 1.00 124.1
AVERTED LOSS
IN OFFSETS
Mahabo 1565 0.80 1252 0.008 0.0040 61 767.0 980.4 0.63 134.5
AVERTED LOSS
IN OFFSETS
Bemangidy 4000 0.90 3600
0.0255
0.0128 61 744.7 1645.7 1.00 901.0
Total All
forest Types
1251.4
Total Fort
Dauphin
Littoral Forest
215.9
Total Littoral
Forest
350.4
66 Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM
APPENDIX 4
Characteristics of High Priority species for which losses and gains
were calculated in Units of Global Distribution (UD)
Higher taxon 2
Scientific name
Non-endemic
species with no
like-for-like offset
Endemism status
IUCN Red
List Status
Unit of measure-
ment
Bespokeloss/gain
measurements
needed?
Global population
size (number
of individuals)
- estimated for
endemic plants
Global
range (hectares)
Amphibians Guibemantis (Mantidactylus) bicalcaratus nov. sp. Regional or local endemic LC
(G. bicalcaratus)
UD 1,595
Amphibians Guibemantis (Mantidactylus) cf pulcher nov. sp. Regional or local endemic NE UD 2,093
Amphibians Guibemantis (Mantidactylus) punctatus nov. sp. Regional or local endemic DD
(G. punctatus)
UD 2,093
Amphibians Madecassophryne truebae Regional endemic EN B1ab(iii) UD 36,373
Amphibians Stumpffia cf tridactyla nov. sp. Regional or local endemic DD UD 2,093
Reptiles Pseudoxyrhopus kely Regional endemic (found only
at two forest sites)
EN B2ab(ii,iii) UD 11,660
Reptiles Phelsuma antanosy Regional endemic (found only
at two forest sites)
CR B2ab(ii,iii,iv) UD Yes 788
Birds Anas melleri Country endemic EN C2a(ii) UD
2,111,126
Birds Ardea humbloti Ye s Country endemic EN C2a(ii) UD 285,554
Terrestrial mammals Microcebus cf rufus nov. sp. Regional or local endemic LC (M. rufus) UD 1,355
Giant pill-millipedes Zoosphaerium alluaudi Regional or local endemic NE UD 890
Giant pill-millipedes Zoosphaerium arborealis Regional or local endemic NE UD 2,093
Giant pill-millipedes Zoosphaerium sp. ‘Sainte-Luce’ Regional or local endemic NE UD 498
Giant pill-millipedes Sphaeromimus inexpectatus Regional or local endemic NE UD 738
Giant pill-millipedes Sphaeromimus splendidus Regional or local endemic NE UD 857
Spirobolid millipedes Riotintobolus mandensis Local endemic NE UD 738
Spirobolid millipedes Riotintobolus minutus Regional or local endemic NE UD 196
Spirobolid millipedes Granitobolus sp.‘black’ Local endemic NE UD 2,093
Spirobolid millipedes Alluviobolus laticlavius Local endemic NE UD 890
Mantises (Mantodea) Nesogalepsus sp. Regional or local endemic NE UD 2,093
Mantises (Mantodea) “Tarachodinae” sp. Regional or local endemic NE UD 2,093
Mantises (Mantodea) Apterocorypha sp. Regional or local endemic NE UD 2,093
Mantises (Mantodea) Platycalymma sp. Regional or local endemic NE UD 2,093
Mantises (Mantodea) Tarachomantis sp. Regional or local endemic NE UD 2,093
Mantises (Mantodea) Tisma freyi Regional or local endemic NE UD 2,093
Mantises (Mantodea) Tisma pauliani Regional or local endemic NE UD 2,093
Mantises (Mantodea) Danuriella irregularis Regional or local endemic NE UD 2,093
Forecasting the path towards a Net Positive Impact on biodiversity for Rio Tinto QMM 67
APPENDIX 4
Characteristics of High Priority species for which losses and gains
were calculated in Units of Global Distribution (UD)
Higher taxon 2
Scientific name
Non-endemic
species with no
like-for-like offset
Endemism status
IUCN Red
List Status
Unit of measure-
ment
Bespokeloss/gain
measurements
needed?
Global population
size (number
of individuals)
- estimated for
endemic plants
Global
range (hectares)
Amphibians Guibemantis (Mantidactylus) bicalcaratus nov. sp. Regional or local endemic LC
(G. bicalcaratus)
UD 1,595
Amphibians Guibemantis (Mantidactylus) cf pulcher nov. sp. Regional or local endemic NE UD 2,093
Amphibians Guibemantis (Mantidactylus) punctatus nov. sp. Regional or local endemic DD
(G. punctatus)
UD 2,093
Amphibians Madecassophryne truebae Regional endemic EN B1ab(iii) UD 36,373
Amphibians Stumpffia cf tridactyla nov. sp. Regional or local endemic DD UD 2,093
Reptiles Pseudoxyrhopus kely Regional endemic (found only
at two forest sites)
EN B2ab(ii,iii) UD 11,660
Reptiles Phelsuma antanosy Regional endemic (found only
at two forest sites)
CR B2ab(ii,iii,iv) UD Yes 788
Birds Anas melleri Country endemic EN C2a(ii) UD
2,111,126
Birds Ardea humbloti Ye s Country endemic EN C2a(ii) UD 285,554
Terrestrial mammals Microcebus cf rufus nov. sp. Regional or local endemic LC (M. rufus) UD 1,355
Giant pill-millipedes Zoosphaerium alluaudi Regional or local endemic NE UD 890
Giant pill-millipedes Zoosphaerium arborealis Regional or local endemic NE UD 2,093
Giant pill-millipedes Zoosphaerium sp. ‘Sainte-Luce’ Regional or local endemic NE UD 498
Giant pill-millipedes Sphaeromimus inexpectatus Regional or local endemic NE UD 738
Giant pill-millipedes Sphaeromimus splendidus Regional or local endemic NE UD 857
Spirobolid millipedes Riotintobolus mandensis Local endemic NE UD 738