SAFRASS-Decision Support Scheme (DSS) to assist the use of river health biomonitoring protocols in Zambia: general aspects, invertebrates (ZISS) and macrophytes (ZMTR) components

Full text

Decision Support Scheme (DSS) to assist the use of river health
biomonitoring protocols in Zambia: general aspects, invertebrates
(ZISS) and macrophytes (ZMTR) components
S.Lowe 1*, H. Dallas 2, M. Kennedy 3, J.C. Taylor 4, C. Gibbins 3,
P. Lang 5, H. Sichingabula 6, K. Saili 6, C. Ntobolo 6., K. Kabangu 6,J. Day 2, F.
Willems 7, J.A. Briggs 1 and K. Murphy 1
1Institute of Biodiversity, Animal Health and Comparative Medicine, University of
Glasgow, Glasgow, Scotland
2Freshwater Research Unit, Department of Biological Sciences,
University of Cape Town, South Africa
3Northern Rivers Institute, University of Aberdeen, Aberdeen, Scotland
4Research Unit for Environmental Science and Management, North-West University,
Potchefstroom, South Africa
5Ecology Assessment Unit, Scottish Environment Protection Agency, East Kilbride,
Scotland
6School of Natural Sciences, University of Zambia, Lusaka, Zambia
7Kasanka Trust, Kasanka National Park, Serenje, Zambia
*corresponding author e-mail: Booley2000@hotmail.com
ACP Science and Technology Programme. A programme of the ACP Group of States, with the
financial assistance of the European Union.
This document has been produced with the financial assistance of the European Union. The contents
of this document are the sole responsibility of the SAFRASS partners and can under no
circumstances be regarded as reflecting the position of the European Union.

Contents
1. Introduction and background
1.1. Aims of SAFRASS biomonitoring protocols:
1.2. Objective of this DSS
1.3. Scope and limitations of the protocols
2. Decision Support Scheme: Narrative
2.1. When to use the aquatic biomonitoring protocols
2.1.1. Routine monitoring sites
2.1.2. To detect point-source pollution
2.1.3. To detect diffuse pollution
2.1.4. Spot sampling or sampling a new monitoring site
2.2. Considerations before sampling:
2.2.1. Prior to departure
2.2.2. At the site
2.2.3. Assessment of a potential site
2.3. Safety
2.4. Site choice:
2.5. During sampling
2.6. Post sampling
2.6.1. Storage of the samples
2.6.2. Metric calculation
2.6.2.1.Invertebrates
2.6.2.2.Macrophytes
2.6.2.3.Diatoms
2.6.3. Data Interpretation
2.6.3.1. General comments
2.6.3.2. Monitoring scenarios
a. Routine monitoring
b. Monitoring point-source pollution
c. Monitoring diffuse pollution
d. Spot sampling or sampling a new monitoring site
3. Decision Support Scheme flow chart
4. References

1. Introduction and background
1.1. Aims of SAFRASS biomonitoring protocols: To detect impairment to tropical river ecosystem
health and water quality using benthic invertebrates, plants and diatoms as bioindicators of water
quality.
1.2. Objective of this Decision Support Scheme (DSS): To assist the practitioner in deciding
when, where and how to use the biomonitoring protocols and to focus attention on issues of safety,
site choice, sampling and data interpretation to ensure the most appropriate and effective
application of the protocols.
1.3. Scope and limitations of the protocols: For use on perennial rivers (or near-perennial rivers)
in Zambia, where access is relatively safe and appropriate for sampling with the equipment
suggested (i.e. <1m water depth in wadeable habitats and, usually, by boat in larger rivers). In their
current form, the protocols are not designed to determine the type of impact causing any detected
impairment of biological quality. However, by careful planning of a monitoring project and
appropriate interpretation of results, the sources and types of pollution may be identified or
inferred. Development of metrics sensitive to particular types of impact such as flow alteration,
sedimentation, or eutrophication (nutrient pollution), may increase the ability of these protocols to
detect specific impacts. Also, it must be reiterated that the protocols are designed to monitor
environmental parameters that, whilst having a strong but indirect influence on human health a nd
well-being, cannot provide a full indication of safe drinking water. Although it is unlikely that sites
with impaired biological quality will provide safe drinking water, sites with high biological quality do
not necessarily indicate safe drinking water. Specific bacteriological, chemical and toxicological tests
are required to provide a guarantee for water of drinking quality. For more on the aims and scope of
the SAFRASS protocols and methodology see Lowe et al. (2013a, b), and Lang et al. (2013).
2. Decision Support Scheme: Narrative
2.1. When to use the aquatic biomonitoring protocols: The following four circumstances
represent the likely scenarios in which the monitoring protocols may be used. A summary of a likely
sampling plan for each scenario is given. Guidelines on interpretation are given towards the end of
this document and in the Methods Manual. Every project has budget, time and staff constraints but
the rule that ‘the more samples the better’ usually applies. Due to typically high variability of the
physical characteristics and biotic communities of rivers, increasing sample number over time (for
routine monitoring, as in the first case below) or in space (for single event sampling plans, as in the
other three cases) will capture more variability and increase confidence in the results and
subsequent interpretation.
2.1.1. Routine monitoring sites: These sites are subjected to regular sampling taken as part of a
medium to long-term project or as part of an organisation’s on-going environmental monitoring
programme. Routinely monitored sites will have data from previous sampling events, making them
very useful for comparison with new samples from the same site, such that deterioration or
improvements in river health and water quality should be readily detectable. To date (2013) in
Zambia, only sites sampled as part of the SAFRASS project or sites near Lusaka sampled by UNZA
have had base-line samples and site data recorded that constitute appropriate sites for historical
comparison (see outputs of SAFRASS project for site localities and associated data). A good choice

for a monitoring site will be in a river used as a water resource by human settlements, of high
conservation priority, and/ or where suspected impacts may arise.
2.1.2. To detect point-source pollution: This is a discrete source of pollution from a single entry
point into the river such as a polluted stream, a storm water/sewage/industrial effluent drain; or an
accidental spill. Setting-up monitoring sites upstream and downstream of potential pollution sources
before any polluting activity is a valuable proactive management strategy. A strategy of reactive
monitoring after a suspected pollution event should include several sites, with at least two or three
sites upstream of the suspected pollution source and several sites spaced appropriately downstream
of the suspected source. The length of river along which to space the monitoring sites will vary
depending on the severity of the pollution and the volume and flow velocity of the river. Ideally,
sites will be located far enough downstream to detect recovery of the biological communities being
monitored.
2.1.3. To detect diffuse pollution: This pollution type occurs over a longer stretch of the river than
point source pollution and is typically a result of activities on a landscape scale in the catchment
rather than a localised activity or accidental spill. Examples of diffuse pollution include impacts from
agriculture, forestry and, in some circumstances, urban development within a catchment. The
impacts of diffuse pollution are generally cumulative down the river so may only be evident in river
reaches lower in the catchment than the source of impact. For example, fine sediment inputs
increasing due to a particular agricultural activity may be most detrimental to aquatic life in reaches
of a river where sediment is deposited on the river bed, and these reaches may be further
downstream than the agricultural area. The spatial scale and type of diffuse pollution may vary
greatly and the number and location of sites will need to be carefully thought-out. Sites upstream,
throughout and extending well downstream of the potential area of pollution source may be
required to detect impacts and locating the source of impact may be difficult.
2.1.4. Spot sampling or sampling a new monitoring site: New sites may be sampled in order to
establish base-line sites for future monitoring or simply as speculative samples to gain some insight
into the environmental health of a river reach. These will be single samples (or a limited number of
samples) taken at sites from which comparable data are not available. In the absence of historical
data from the site, establishing reference sites for comparison on unimpacted rivers close by which
have similar hydrology (catchment size, flow regime, discharge, stream width) and habitat
characteristics (substrate type, canopy cover etc) will provide the best indication of the biological
community structure, and therefore biomonitoring metric scores, to expect at the site.
Interpretation of the ecological status at such isolated sites is generally difficult and it is ill-advised to
base conclusions or make management recommendations concerning the health of a river reach
based on an isolated sample. However, guidelines are given below and in the Methodology Manual
(Lowe et al 2013b).
2.2. Considerations before sampling:
2.2.1. Prior to departure:
If details of the site are not known prior to departure, find the site(s) on a map or Google
Earth/Maps to determine potential size of the river and access to the river. If the site is on a large
river (>20m water surface width), a boat may need to be sourced either prior to the trip or once on
site. If the site is on private land or a nature reserve permits or permission will be required and even

if on public land, it may be advisable to inform the local chief or headman about the project and seek
consent to sample, especially if a village is nearby.
Write a checklist of all the equipment and ensure it is operational and packed prior to departure. See
Methodology Manual (Lowe et al., 2013b) for required and suggested equipment.
Establish whether the site has been sampled previously with biotic and physical characteristics
recorded (e.g. possibly as part of the SAFRASS project). If so, find information on where, when and
how the sampling was done. This will direct your sampling effort and assist with interpretation of
data.
If not, fill-in a new site record sheet at the site (see Methodology Manual Appendix 4: Lowe et al,
2013b).
2.2.2. At the site:
Does the river flow throughout the year for most years? (Note: ask as many sources as possible.
Local communities are often the best people to ask.)
If yes, proceed.
The river is perennial and suitable for the biomonitoring protocols. If the river stops flowing in
drought years only, the river is still suitable for monitoring, but make a note of this).
If no, does the river contain water all year and only stop flowing for 2 months or less (e.g. in October
and November)?
If yes, proceed.
The river is marginally intermittent and suitable for application of monitoring techniques but with
caution for interpretation of results (see later).
If no, the river reach is intermittent or ephemeral and the site is not suitable for the application of
the protocols
2.2.3. Assessment of a potential site:
Always take time to walk upstream and downstream (preferably on both banks) as knowledge of a
site and its surrounding area is invaluable and always yields insight to help direct sampling,
interpretation and general understanding of river systems. Whenever possible, talk to local people
or guides/rangers/managers (e.g. ZAWA) to gain information on safety and river characteristics (e.g.
perenniality, high flow/low flow level, resource use and potential impacts etc).
2.3. Safety:
a) Are there dangerous animals in the river?
Ask local people, look for tracks and use intuition.
If yes (or suspect yes), proceed with extreme caution or, if the threat is considered very high,
abandon sampling until an experienced (and preferably armed) scout is available or, if on a larger
river, a boat can be acquired. If you proceed, ensure colleagues are looking whilst you are sampling
and spend a minimum time at the water edge. Set up field equipment away from the water edge.

If you don’t know, assume there are dangerous animals and act accordingly.
If confirmed that no dangerous animals are present, still proceed with caution and always be aware
of potential danger.
b) Are there water-borne diseases, such as Schistosomiasis, prevalent?
Local people are generally knowledgeable about these issues.
If yes, extra care should be taken. Use rubber gloves (preferably elbow length), waders and wash
potentially contaminated body parts with ethanol or antiseptic (e.g. Dettol) solution after sampling.
If no, still proceed with caution, especially if close to, or downstream of, human settlements. Always
wash hands and any areas of broken skin with antiseptic wipes.
2.4. Site choice:
Available and accessible habitats:
a) Are there localised impacts that are not part of the monitoring concern? For example, is
there a road bridge (which will normally cause shading, change channel shape and have altered
habitat and substrate) or an instream obstruction or channel alteration which may change flow or
habitat?
If yes, try to find a site away from the local impact that is more representative of the natural stream
condition or more similar to the rest of the stream reach, preferably upstream of the impact. If you
cannot find a site that is accessible away from the impact, then sample the habitat around the
impact, making a note of the impacts (on substrate, flow etc) to assist with interpretation.
If there is no impact, proceed with sampling at the most convenient and safest access point that
allows access to all available habitats. This may not be the closest point to the vehicle/point of
arrival.
b) Are suitable substrates available for diatom sampling? (i.e. semi-permanent, moveable
substrates such as cobbles (medium-sized stones) or reeds).
c) What macrophyte growth-types are present and accessible, both for macrophyte sampling
and as invertebrate habitat? (marginal, emergent, submerged, floating)
d) Are suitable benthic substrates and flow-types present and accessible for sampling aquatic
invertebrates? ( See the site record sheet for habitat types)
e) Can you safely access these habitats for sampling?
If yes, sample these habitats in approximate proportion to their availability. For example, for ZISS
sampling, if 75% (three quarters) of the stones-in-current biotope are cobbles in a run and 25% are
boulders in a riffle, spend about 90 seconds sampling in the run and 30 seconds in the riffle.
2.5. During sampling:
How and in what order should the different organism groups be sampled?

See the SAFRASS Methodology Manual, training video, and instructions on the sampling data-
recording sheets for details of how to sample.
The sampling should be carried out in order of least disturbance: usually diatoms first, then
macrophytes (in practice diatoms and macrophytes can often be sampled at the same time), then
invertebrates. The collector for each component should discuss their sampling ‘plan of action’ with
the other collectors (if different) to coordinate and cause least disturbance to areas the others wish
to sample. Most often each component can be collected in conjunction: the invertebrate collector
can pick stones for diatom collection and plants for the botanist, the diatomist can pick invertebrates
from the stones from which diatoms are scrubbed and the botanist can do the same from plant
samples.
2.6. Post sampling and data interpretation
2.6.1. Storage of the sample: Does the sample need to be collected and preserved for future
analysis?
It is always advisable to keep a few representative specimens of all taxa collected from each site or
cluster of sites. These specimens are useful reference material for museum records, as proof of
correct family identification, or in order to confirm the identification of uncertain taxa. Animals can
be stored in small vials (10ml tubes or 40ml specimen bottles) with preservative (~80% ethanol).
These will be useful for medium-term storage only as evaporation will occur.
The whole invertebrate sample should only be preserved if the project has a specific requirement to
do so. Unless there are dedicated researchers with allocated time, it is unlikely that samples will be
sorted and identified (i.e. to genus or species level). Unnecessary storing and processing wastes
laboratory time and space.
Plants can be dried in a plant press using paper towel or other absorbent paper and in the longer
term stored flat, mounted on sheets of acid free paper as herbarium specimens: see Methodology
Manual for more detail (Lowe et al., 2013b).
Diatom procedures are reported separately: see Lang et al. (2013) “Proposed procedure for the
sampling, preparation and analysis of benthic diatoms from Zambian rivers: a bioassessment and
decision support tool applicable to freshwater ecoregions in tropical southern Africa”.
2.6.2. Metric calculation
For more detailed notes on metrics derived from the ZISS protocol and interpretation guidelines, see
the Methodology Manual (Lowe et al., 2013b). The following is a brief guideline to interpreting data
collected as part of one of the four scenarios described above (i-iv in Section 2.1).
2.6.2.1 Invertebrates: The three core metrics for ZISS data are: 1) the number of ZISS taxa (families)
recorded at a site; 2) the total ZISS score (sum of sensitivity scores for each taxa recorded); and 3)
the Average Score Per Taxon (ASPT) (ZISS score divided by number of ZISS taxa). These values should
be calculated for each habitat sampled at each site. Schematic diagrams of hypothetical sampling
projects representing different monitoring scenarios and associated graphs of metric values are
presented to assist project planning and data interpretation (Fig 1a-c and Fig 2a-c).

2.6.2.2. Macrophytes: Using the protocol described in the Methodology Manual (Lowe et al.,
2013b), and the species trophic ranking scores (TRS) in the SAFRASS macrophyte picture
identification guide (Kennedy & Murphy, 2012), the Zambian Macrophyte Trophic Ranking (ZMTR)
score for a site can be calculated by summing the TRS score for each macrophyte species present,
and dividing this by the overall number of species.
2.6.2.3. Diatoms: Diatom procedures are reported separately: see Lang et al. (2013) “Proposed
procedure for the sampling, preparation and analysis of benthic diatoms from Zambian rivers: a
bioassessment and decision support tool applicable to freshwater ecoregions in tropical southern
Africa”.
2.6.3. Data interpretation
2.6.3.1 General comments
Several biomonitoring systems exist, worldwide, that use statistical procedures to greatly enhance
the confidence of interpreting results from a monitoring sample. Some monitoring systems use a
predictive approach based on measuring specific physical characteristics of the river to predict the
expected biotic assemblage (or probability of species occurrence) as a comparison to the observed
assemblage. Examples of such systems include RIVPACS in the UK and AUSRIVAS in Australia.
Development of such a system is beyond the scope of the SAFRASS pilot project as the number of
reference sites required for an area as large as Zambia would be many hundreds. Thus, SAFRASS has
taken a simpler approach that is appropriate to circumstances in Zambia.
The basis for interpretation of metric values for a particular sample is understanding the ‘natural’
variability of metric values from the site or similar sites (i.e. variability of reference sites) and
whether the metric value for a new sample falls within (unimpacted) or outside (impacted) of that
variation. For example, if an unimpacted site has previously been sampled four times with ASPT
values of 6.6, 7.0, 6.8 and 6.8, then a value of 6.0 obtained from a new sample may be a cause for
concern. However, comparison between only one previous sample with an ASPT value of 6.6 with a
new value of 6.0 is very difficult to interpret as there is no appreciation of the expected variance in
ASPT at that site.
Although the notes below provide interpretation guide-lines to users with limited experience, it
should be noted that there can be no substitute for experience in order to derive maximum
interpretive value from biomonitoring data. Consideration of abundance of individual taxa in relation
to metric scores (the current metric scores used in ZISS are only sensitive to presence/absence), how
metric scores relate to each other (for example if ASPT is unchanged but number of taxa and total
ZISS score changes greatly) and knowledge of how the biology of individual taxa may relate to site-
specific physical characters and to differences in score or abundance between samples are all
important factors that require understanding of river ecology in order to fully interpret data.
2.6.3.2. Monitoring scenarios:
The most effective use of biomonitoring protocols is to sample sites as part of a planned monitoring
project. These projects typically feature three main types of monitoring scenarios as described
above in section 2.1. Brief explanations for interpreting data from each of these scenarios are given
below. In addition, hypothetical examples for each of these scenarios using invertebrates (ZISS) are

schematically represented in Figures 1a-c, with hypothetical results graphically represented in Figure
2a-c. These hypothetical projects are presented by way of example to assist project planning and
interpretation of results and, although the invertebrate metric ASPT is given by way of example, the
principles of interpretation are the same for any invertebrate, diatom or macrophyte metric
developed during SAFRASS. For real-life examples of invertebrate monitoring to detect impacts of
human activities on river health see Lowe et al. (2013d): section 4.1. Each of the hypothetical
scenarios has been limited to six sample years (scenario a) or six sample sites (scenarios b and c).
Increased number of sample sites in reference reaches and reaches of potential impact are always
desirable, but time and budget constraints usually limit the number of samples that can be collected.
a) Routine monitoring (sampling a site repeatedly over time, preferably at regular intervals):
Metric values from samples of each of the biotic indices (macrophytes, diatoms and
macroinvertebrates) can be compared to those from previous samples taken at the monitoring site.
If data are available from many previous sampling events at a site, then comparison of the metrics of
any new sample from that site can be made with increased confidence. Ideally, routinely monitored
sites will be located at sites of high conservation or human health value or sites of particular
pollution concern. Spatial comparison is also useful in these instances, as shown in the hypothetical
example (Fig 1a). In the example, the first three years show little change in ASPT at both sites 1 and
2. In year 4, site 2, but not site 1, shows a large decrease in ASPT relative to the first three years,
with a gradual recovery close to initial values by year 6. A likely interpretation of these results is that
somewhere between sites 1 and 2 a pollution incident occurred between the time of sampling in
year 3 and the time of sampling in year 4. This pollution event compromised ecosystem health which
appeared to persist for nearly two years.
b) Monitoring point-source pollution (sampling several sites above and below a discrete source of
pollution entering directly into the river):
Comparison between sites can be made along a river profile from upstream to downstream.
Typically, assuming sites are in similar river types, one would expect that sites upstream of the
pollution source would have similar values for each metric (and possibly similar taxa). Sites
immediately downstream of a pollution source can be expected to have reduced metric values
relative to those upstream. Metric values can be expected to gradually recover to values similar to
those recorded from sites upstream of the pollution source which reflects dilution, degradation and
uptake (biological processing) of the pollutant. The distance over which recovery takes place (i.e.
length of river impacted by the pollution) will be dependent on factors such as the severity and type
of pollution and the size, hydrological, chemical and biological characteristics of the river. The
second hypothetical scenario (Figure 1b) exemplifies this type of monitoring project: sites 1, 2 and 3
have similar ASPT values (Figure 2b), but the value for site 4 is much reduced with recovery of ASPT
to values close to upstream sites by site 6. These results imply that a pollution event has occurred
recently or is on-going somewhere between sites 3 and 4. The pollution negatively impacts
ecosystem health for a distance of at least that between sites 4 to 6.
c) Monitoring diffuse pollution (sampling several sites on the river above, within and below a source
of pollution occurring over a larger area in the catchment):

As with point-source pollution, comparison between sites can be made along a river profile from
upstream to downstream or with sites on other rivers that are physically (and biologically) similar
but without pollution. The difference between diffuse and discrete pollution sources is that with
diffuse pollution, the pollutant or agent of environmental degradation enters the river over a longer
reach and the impacts are likely to be cumulative which may be more evident with increasing
distance downstream of the source of impact. As such, the profile of change in metric values for
samples taken from sites upstream to downstream of the suspected impact are likely to be different
from point-source pollution. Typically, a gradual reduction in metric values may be detected from
sites upstream of the area of the impact (e.g. from land-use change) progressively downstream. A
gradual recovery of metric values may or may not be evident, depending on the length of the river
sampled and the extent of the impact. Diffuse pollution associated with land-use change may take
place over a large area in a catchment and a long stretch of river. As such, physical characteristics
and biological communities of reference sites upstream of the impacted area may be naturally
different from those below the area of impact. Also, additional sources of impact (diffuse or point-
source) may influence river health over such a long stretch of river. This increases the difficulty of
interpreting metric values unless previous data is available from the river reach under investigation
or from comparable reaches on similar nearby unimpacted rivers. In the hypothetical scenario of
diffuse pollution monitoring (Fig. 1c), there is a gradual decline of ASPT values downstream (Figure
2c), the most dramatic of which occur outside of the suspected source of pollution. This implies
accumulation of the impact and is typical of sedimentation which tends to accumulate in reaches
that are slower flowing and have reduced erosive or scouring properties. In the absence of other
sources of pollution, the hypothetical results presented in scenario c can be interpreted as land-use
in the catchment above site 5 and below site 1 having a negative impact on river health whic h
increases downstream for an unknown distance. Follow-up sampling to this project would include
extending sampling downstream to establish if remediation of impacts occurs and sampling of an
equivalent river or rivers to assess natural patterns of biotic change (specifically, ASPT) over the
length of river under consideration in order to compare such changes with potential impacts from
the suspected diffuse pollution source.
d) Spot sampling or sampling a new monitoring site:
Due to the extensive geographic coverage of SAFRASS sampling sites, the low density of sites does
not allow for development of an accurate predictive biomonitoring approach. As such, it is difficult
to determine ecosystem health of individual sites monitored outwith a project such as those for
temporal (longer-term) or spatial (upstream to downstream) monitoring suggested above. One way
of overcoming the problem of establishing impact is by comparison to samples from unimpacted
sites on similar river reaches with similar habitat in order to establish an expected range of metric
scores. As a guideline to this method, Table 1 summarises values of the average (mean) and
standard deviation (measure of variance) of the three ZISS metrics calculated for invertebrate
samples from the three habitats (vegetation, stones and GSM) at unimpacted sites. Sites were
grouped by freshwater ecoregion (FE) and by ‘high’ order (7-9: “downstream”) and ‘low’ order (3-6:
“upstream”) river reaches. It should be noted, however, that within these groups (or ‘river types’)
there is high variability of physical characteristics and multivariate analyses (data not shown)
indicate only a weak association between these predetermined river types and biological community
structure. Only groups with three or more samples are shown, such that vegetation samples from
both stream order categories in all five ecoregions are represented, but that only low stream order

river reaches in three or four ecoregions are represented for stone samples and GSM samples
respectively.
Number of taxa and total ZISS score from vegetation samples are generally lower in larger, high
order rivers than in smaller, low order rivers, particularly from Bangweulu-Mweru and Kafue FEs.
Within the two stream order groups, metric values for vegetation samples are relatively consistent
between ecoregions with the exception of lower ASPT values from low stream order reaches in
Kafue and Upper Zambezi floodplain ecoregions.
Benthic samples (stones and GSM) from unimpacted high order river reaches were rare or absent
due to a combination of difficulty of access to the river bed in larger rivers and the fact that many of
these reaches are classified as impacted, often due to flow regulation by impoundments upstream.
Mean average metric values from stone samples are consistent between the three ecoregions
represented, but the variance (as measured by standard deviation) for all values are consistently
high indicating that further refinement of site grouping (increased number of river types) is required
to more confidently apply reference ranges for metric values. Combining values from all ecoregions
shows similar metric values between large and small rivers for stone samples. As for stones samples,
mean average metric values from GSM samples are consistent between ecoregions and combining
values from all ecoregions shows similar metric values between large and small rivers. Values for all
metrics for GSM samples are, however, consistently much lower, with lower variance, than for stone
samples.
In conclusion (for application of the most appropriate reference values derived from SAFRASS data
to interpret the impact status of a new sample):
The most appropriate reference value of metrics for stone samples and GSM samples is a single
mean average and standard deviation value for each habitat from all unimpacted sites (where these
habitats are available) across the country (Table 1: values in bold).
Metric values for vegetation samples should be compared within ecoregion and stream order
groups. Stream order can be derived from the SAFRASS sampling sites map (Lowe et al., 2013a:
Appendix 1). Alternatively, high stream order (order 7-9) is a close proxy for wide rivers (>20m water
surface width) and low stream order is a close proxy for narrower rivers (<20m water surface width)
so measuring water surface width at a site will indicate which reference group is appropriate.
The high variance (standard deviation) within these groups for each habitat type indicates that
further sampling is required to confidently assign reference values to more refined groups of river
types.

Table 1. Mean and standard deviation (in parentheses) of ZISS metrics for invertebrate samples from
unimpacted SAFRASS sampling sites (sampled during 2010 – 2012) grouped by ecoregion and two
stream order categories, where “high” = order 7 – 9, and “low” = order 3 – 6. Numbers in
parentheses following stream order category for combined ecoregion values are the number of sites
used.
habitat
ecoregion
stream
order
ZISS Score
# Taxa
ASPT
vegetation
Bangweulu_Mweru
high
42 (20)
8.3 (4.2)
5.1 (0.6)
vegetation
Bangweulu_Mweru
low
80.9 (28)
13.6 (4.3)
5.9 (0.7)
vegetation
Kafue
high
53.6 (9)
8.6 (1.5)
6.3 (0.7)
vegetation
Kafue
low
75 (21.4)
15.2 (3.8)
4.9 (0.5)
vegetation
Mid-Zambezi_Luangwa
high
81 (23.4)
14 (5.3)
5.9 (0.6)
vegetation
Mid-Zambezi_Luangwa
low
75.5 (20.6)
13 (3)
5.8 (1)
vegetation
Upper Zambezi floodplain
high
60.7 (15.4)
11 (3.5)
5.6 (0.5)
vegetation
Upper Zambezi floodplain
low
64.8 (8.2)
13.8 (1.5)
4.7 (0.7)
vegetation
Zambezi headwaters
high
70.3 (8.6)
11.7 (1.5)
6 (0.1)
vegetation
Zambezi headwaters
low
90.8 (31.5)
15.8 (5)
5.8 (1.1)
vegetation
All ecoregions
high (17)
60.6 (18.8)
10.5 (3.6)
5.8 (0.6)
vegetation
All ecoregions
low (77)
81.2 (27.1)
14.2 (4.2)
5.7 (0.9)
vegetation
All ecoregions
all (94)
77.5 (27)
13.5 (4.3)
5.7 (0.9)
stones
Bangweulu_Mweru
low
99 (41)
14 (5.8)
7.2 (1.1)
stones
Mid-Zambezi_Luangwa
low
100.2 (40.1)
15 (5.3)
6.7 (1.2)
stones
Zambezi headwaters
low
106.2 (44.5)
16 (4.9)
6.5 (1.1)
stones
All ecoregions
high (7)
102.9 (26)
14.7 (3.6)
7 (0.6)
stones
All ecoregions
low (41)
98.9 (38.7)
14.5 (5.3)
6.9 (1.1)
stones
All ecoregions
all (48)
99.4 (36.9)
14.5 (5.1)
6.9 (1)
GSM
Bangweulu_Mweru
low
37.2 (4.3)
7 (1)
5.4 (0.9)
GSM
Kafue
low
48 (13.8)
10 (2.7)
4.8 (0.2)
GSM
Mid-Zambezi_Luangwa
low
41.3 (35.1)
7.8 (5.1)
5 (0.9)
GSM
Zambezi headwaters
low
49.4 (39.8)
8.7 (4.8)
5.3 (1.4)
GSM
All ecoregions
high (4)
43.5 (14.2)
8.3 (1.5)
5.2 (1.2)
GSM
All ecoregions
low (21)
41.9 (25.8)
8.1 (3.8)
5 (0.9)
GSM
All ecoregions
all (25)
42.1 (24.1)
8.1 (3.5)
5 (0.9)

Figure 1. Schematic maps of typical monitoring scenarios (hypothetical examples)
Figure2. Hypothetical results for one metric (ASPT) from each of the typical monitoring scenarios
4
4.5
5
5.5
6
6.5
7
7.5
1 2 3 4 5 6
ASPT
year
a) Routine monitoring of
two sites site 1
site 2
4
4.5
5
5.5
6
6.5
7
7.5
1 2 3 4 5 6
ASPT
site number
b) Point-source pollution
4
4.5
5
5.5
6
6.5
7
7.5
123456
ASPT
site number
c) diffuse pollution

3. Decision Support Scheme Chart
Decision Support Scheme
1) Prior to departure
Context of
sampling
•Routine monitoring (over time)
•Point source pollution monitoring
•Diffuse pollution monitoring
Additional
considerations Do you have the following?
Historic data
•Checklist of equipment
•Appropriate permits (esp inside NPs but useful in all situations)
•Permission to access river if required
•Vehicle appropriate to potential conditions (assume the worst case scenario)
•Data sheets
Yes.
Acquire the historic data in order to:
•locate the site, river access points and specific habitats to sample;
•derive constant variables for site characterisation survey;
•assist interpretation of new results (how biota and/or physical
variables have changed over time)
Has the site previously
been sampled using
biomonitoring protocols
(e.g. During SAFRASS
project)? No.
Locate reach on a digital or hard map to
establish potential site access
Yes.
Plan site sampling in conjunction
with other sites to be sampled
for the project
No.
Sampling not advised
(see DSS narative)
Is the planned sampling
site(s) part of a structured
monitoring project?
Decision Support Scheme
2) At the site
Perenniality
Mode of
sampling
Is water surface width
greater than 20m?
(If possible, check
before departure using
previous sampling data
or hard/digital map)
Safety
Yes.
Can benthic substrate be
sampled? (i.e. is substrate
within ~1m and in flow
that is safe to stand in)
Yes.
Proceed with extreme
caution. See DSS
narrative for advice
Safety 1. Are there
dangerous animals
in the river?
No.
Proceed but maintain
awareness of potential danger
Yes.
Continue sampling
No.
Does the river contain water all year
and only stop flowing for 2 months
or less (e.g. October and November)
Does the river flow
throughout the year in
years of normal rainfall?
Yes.
Continue sampling
No.
SAFRASS protocols are not
appropriate as the river reach is
intermittent or ephemeral
Not sure.
Assume there are
dangerous animals. See
DSS narrative for advice
Yes.
Use long rubber gloves and waders.
Wash potentially contaminated
body parts with ethanol or
antiseptic solution after sampling
Safety 2. Are
there water-
borne
diseases in
the river?
No.
Proceed but maintain awareness
of potential danger
Not sure.
Assume there are diseases
especially if close to settlements
No.
Proceed with sampling as for
smaller rivers (i.e. from bank-
side or, if depth and flow
allow, within the channel)
Yes.
Proceed with sampling as for smaller rivers
No.
Is the river safely
navigable and is a
boat available?
Yes.
Proceed with sampling as
for larger rivers by boat
No.
Proceed with macrophyte and
diatom sampling if possible,
but abandon ZISS

Site selection
Site characterisation
(recording physical
variables)
Biotic sampling
Decision Support Scheme
3) At the site (cont.)
No.
Are there localised impacts
at the potential site that are
not part of the monitoring
concern? (e.g. road bridge)
Yes.
Try to find a site typical of the stream
reach, preferably upstream of the impact.
If there is no accessible site away from the
impact, then sample close to the impact,
recording the type and severity of impact
on the New Site Characterisation Sheet
No.
Find a site typical of the stream reach.
Proceed with sampling at the most
convenient and safest access point that
allows access to all available habitats.
This may not be the closest point to the
vehicle/point of arrival
Has the site previously
been sampled using
SAFRASS protocols and
is site characterisation
data available?
Yes.
Use data from previous
samples (coordinates, site
photo and site drawing) to
sample the same habitats
1) Are suitable substrates available and accessible for
diatom sampling?
2) Are marginal or aquatic macrophytes present and either
identifiable from the bank-side or retrievable?
3) Are benthic substrates and or macrophyte habitat
present and accessible for sampling aquatic invertebrates?
Yes.
Proceed with sampling available biota
as detailed in the SAFRASS Methods
Manual (WP 12)
Note: if macrophytes are present, all
three biotic groups can be sampled
No.
The site is not suitable for biomonitoring.
Interpretation of site impacts are
stronger with increased number of biotic
indices that are sampled.
No.
Complete the New Site
Characterisation Sheet
Has the site previously
been sampled using
SAFRASS protocols and
is site characterisation
data available?
Yes.
No need to complete a new site
characterisation sheet, but the user must
record ZISS habitats sampled and any
variable that may change between season or
year and measure water chemistry variables

4. References
Kennedy M.P. & Murphy K.J. (2012). A picture guide to aquatic plants of Zambian rivers. SAFRASS
Deliverable Report to the African, Caribbean and Pacific Group of States (ACP Group) Science and
Technology Programme, Contract No. AFS/2009/219013 .University of Aberdeen. 25 pp.
Lang P., Taylor J.C., Bertolli L., Lowe S., Dallas H., Kennedy M.P., Gibbins C., Sichingabula H., Saili K.,
Day J., Willems F., Briggs J.A., & Murphy K.J. (2013). Proposed procedure for the sampling,
preparation and analysis of benthic diatoms from Zambian rivers: a bioassessment and decision
support tool applicable to freshwater ecoregions in tropical southern Africa. SAFRASS Deliverable
Report to the African, Caribbean and Pacific Group of States (ACP Group) Science and Technology
Programme, Contract No. AFS/2009/219013 . Ecology Assessment Unit, Scottish Environment
Protection Agency, East Kilbride, Scotland. 11 pp.
Lowe S., Dallas H., Kennedy M.P., Taylor J.C., Gibbins C., Lang P., Sichingabula H., Saili K., Ntobolo C.,
Kabangu K., Day J., Willems F., Briggs J.A., & Murphy K.J. (2013a). The SAFRASS biomonitoring
scheme: general aspects, macrophytes (ZMTR) and benthic macroinvertebrates (ZISS) protocols.
SAFRASS Deliverable Report to the African, Caribbean and Pacific Group of States (ACP Group)
Science and Technology Programme, Contract No. AFS/2009/219013 . University of Glasgow,
Glasgow, Scotland. 20 pp.
Lowe S., Dallas H., Kennedy M.P., Taylor J.C., Gibbins C., Lang P., Sichingabula H., Saili K., Ntobolo C.,
Kabangu K., Day J., Willems F., Briggs J.A., & Murphy K.J. (2013b). SAFRASS Methodology Manual.
SAFRASS Deliverable Report to the African, Caribbean and Pacific Group of States (ACP Group)
Science and Technology Programme, Contract No. AFS/2009/219013 . University of Glasgow,
Glasgow, Scotland. 36 pp.
Lowe S., Dallas H., Kennedy M.P., Taylor J.C., Gibbins C., Lang P., Sichingabula H., Saili K., Ntobolo C.,
Kabangu K., Day J., Willems F., Briggs J.A., & Murphy K.J. (2013d). Assessment of performance of the
pilot river biomonitoring scheme. SAFRASS Deliverable Report to the African, Caribbean and Pacific
Group of States (ACP Group) Science and Technology Programme, Contract No. AFS/2009/219013 .
University of Glasgow, Glasgow, Scotland. 26 pp.