Content uploaded by José Lopez
Author content
All content in this area was uploaded by José Lopez on Apr 10, 2023
Content may be subject to copyright.
ORIGINAL PAPER
Disturbance and forest recovery in the Monarch Butterfly
Biosphere Reserve, Mexico
Jose
´Lo
´pez-Garcı
´a
1
•Rafael M. Navarro-Cerrillo
2
Received: 29 August 2018 / Accepted: 8 January 2019
ÓNortheast Forestry University 2019
Abstract Analysis of the change in forest cover is
important to determine stand dynamics and the processes
involved in disturbance and recovery. Forests of the core
zone of the Monarch Butterfly Biosphere Reserve were
studied using photo interpretation techniques, considering
biennial changes between 1999 and 2013 and changes over
the whole period of study 1999–2013. Error matrices were
elaborated to determine the processes of change involved
in both recovery and disturbance. The biennial changes for
the whole period amounted to 2274 ha; 343 ha accounted
as degraded in more than one biennial period. The total
changes in forest cover between 1999 and 2013 involved
4902 ha, out of which, 2912 ha were affected by distur-
bance, and 1990 ha were recovered. For density and 2013
vegetation cover maps, the overall accuracy was 95.6% and
90.2%, respectively. By combining both maps, gradual
processes were revealed that were not evident in separate
analyses. This methodology is useful for the management
and conservation of natural protected areas.
Keywords Illegal logging Forest density change Forest
recovery Conifer forests Monarch Butterfly Biosphere
Reserve
Introduction
Forests are major reservoirs of biodiversity, frequently
important as carbon (C) sinks, and provide important
ecosystem services (Anderson-Teixeira and DeLucia
2011). Inter-decadal variability in the rate of atmospheric
CO
2
shows changes in net land carbon sink, with estimates
of 0.3 ±0.9, 1.0 ±0.6 and 0.9 ±0.6 Gt C/year for the
1980s, 1990s and 2000–2005 periods, respectively. A
combining of techniques gives an estimate of the flux of
CO
2
to the atmosphere from land use changes of 1.6 (0.5 to
2.7) Gt C year
-1
for the 1990 s. The global sequestration of
C by forests between 1990 and 2007 has been estimated at
2.4 ±0.4 Picograms (Pg) C a
-1
. Geographically,
471 ±93 Pg C (54.6%) is stored in tropical forests,
272 ±23 Pg C (31.5%) in boreal, and 119 ±6PgC
(13.8%) in temperate forests. The C stock density of tem-
perate forests is 155 Mg C ha
-1
. The capture of C in
temperate forests increased by 17% from 2000 to 2007,
compared to the period 1990-1999 (Pan et al. 2011).
Forest regeneration influences the global carbon cycle,
and understanding how climate changes will affect regen-
eration and carbon storage is necessary to predict the rates
of increase in atmospheric CO
2
in the coming decades
(Miller et al. 2016). The forested areas of Mexico
decreased from 69.76 Million hectares (Mha) to 66.04 Mha
between 1990 and 2015; protected forest areas increased
Corresponding editor: Zhu Hong.
Project funding: This work was funded by CONACYT and DGAPA
within the sabbatical abroad program for the consolidation of research
groups ‘‘Dynamics of deforestation, forest degradation and recovery
in the Monarch Butterfly Biosphere Reserve’’.
The online version is available at http://www.springerlink.com.
&Jose
´Lo
´pez-Garcı
´a
jlopez@unam.mx
1
Instituto de Geografı
´a, Universidad Nacional Auto
´noma de
Me
´xico, Ciudad Universitaria, C.P. 04510 Coyoaca
´n, DF,
Mexico
2
Departamento de Ingenierı
´a Forestal, Universidad de
Co
´rdoba, Campus de Rabanales, Carretera Nacional IV km
396, 14014 Co
´rdoba, Spain
123
J. For. Res.
https://doi.org/10.1007/s11676-019-00964-3
from 3.78 to 8.8 Mha from 2000 to 2015 (FAO 2015).
Mexico suffered a net forest loss of 0.35 Mha/year between
2000 and 2010 (FAO 2010). Forest degradation is con-
sidered to be the reduction of the capacity of forests to
provide goods and services (FAO 2010), and a decrease in
the density of trees per hectare may be partially considered
as forest degradation, and in contrast, an increase in the
density of trees per hectare may be considered partially as
forest improvement. These changes in forest cover, even if
temporary, modify the biomass of the ecosystem. There-
fore, assessment of cover change is important in order to
understand how forest C stocks change, since decreases are
often reflected in biomass reduction and in canopy and/or
tree coverage per unit area (Lund 2014).
Reduction of CO
2
emissions due to deforestation and
forest degradation (REDD) is an initiative United Nations
to create financial value for carbon storage by offering
incentives to developing country farmers to reduce emis-
sions from forested land and to invest in programs of low C
consumption towards sustainable development. It is,
therefore, a mitigation mechanism that arises as a result of
the avoidance of deforestation. For REDD to become
internationally accepted, it must meet at least three criteria:
reduce real emissions by being effective; minimize costs by
being efficient; and, reduce unwanted social and ecological
commitments by being equitable and providing co-benefits
(Angelsen and Wertz-Kanounnikoff 2008). The
REDD ?program goes beyond REDD’s reforestation and
forest degradation and includes the roles of conservation,
sustainable forest management and the improvement of
forest C stocks. However, in general, changes in forest
cover density have not been analysed in studies of forest
change, perhaps because of its qualitative nature. This
allows for the determination of changes in tree density,
generally in temperate forests of Natural Protected Areas
(NPA), as in the case of the Monarch Butterfly Biosphere
Reserve (Brower et al. 2002;Lo
´pez-Garcı
´a et al.
2014,2016; Manzo-Delgado et al. 2014; Vidal et al. 2014).
The states of Me
´xico and Michoaca
´n have experienced
significant changes in land cover since the identification of
the monarch butterfly sanctuaries in 1975 (Ramı
´rez et al.
2003). Annual deforestation increased from 1.7% from
1971 to 1984 to a maximum of 2.4% from 1984 to 1999
(Brower et al. 2002). Additionally, the discontent of the
local population owed to the designation of this Natural
Protected Area in 1986, on which a protection zone of
116,110 ha, was established, including five buffer and six
core zones, caused the reaction of the people to cut and
burn the forest of the Chivati-Huacal hills influencing the
disappearance of the colony of monarch butterflies from
this place. Under these circumstances, the main threat to
the hibernation of monarch butterflies in Mexico has been
deforestation and forest degradation due to illegal logging
(Brower et al. 2012). Illegal logging is still the main
problem facing the Reserve (Honey-Rose
´s2009).
More recently, a study of the change in forest cover,
with data from 1971, 1984 and 1999, was carried out for an
area of 40,000 ha. This was the basis for the expansion of
the NPA and its designation as the Monarch Butterfly
Biosphere Reserve (MBBR) (Brower et al. 2002). Studies
also have compared changes in forest cover for 1993 to
2006 (Navarrete et al. 2012) with changes in the period
2006–2010 (Champo-Jime
´nez et al. 2012). Digital aerial
photography has also been used to evaluate cover changes
in the core zone in 2003–2009 (Lo
´pez-Garcı
´a2011). A
similar study was carried out in the core area over the
period 2001–2012 to assess landscape changes related to
small and large-scale felling of trees and to natural events
(Vidal et al. 2014). In general, studies to determine the
effectiveness of protected areas indicate that the MBBR
increased forest cover between 1993 and 2002 (Figueroa
et al. 2011).
However, changes in forest cover in the Reserve require
a better distinction between deforestation and forest
degradation, and between direct and indirect factors, as
well as clarification of the causes of these processes.
Deforestation is defined as ‘‘the conversion of forests to
other land use or the long-term reduction of canopy cover
below the minimum threshold of 10%’’, and forest degra-
dation as ‘‘changes within the forest that negatively affect
the structure or function of the site’’ (FAO 2000). Forest
degradation is generally associated with a significant
reduction of forest cover (Lund 2014).
Photo interpretation allows for a description of the
natural dynamics of forest degradation and deforestation
through the evaluation of forest cover as an indicator of the
degree of disturbance (Brower et al. 2002;Lo
´pez-Garcı
´a
2011;Lo
´pez-Garcı
´a et al. 2014; Manzo-Delgado et al.
2014; Vidal et al. 2014). In addition, aerial photography
allows for the evaluation of the vertical dimension of
vegetation and the use of longer time records to evaluate
changes in vegetation (Fensham and Fairfax 2002). These
advantages justify the additional use of photo interpretation
in studies of vegetation change (Dahdouh-Guebas et al.
2000; Fensham and Fairfax 2002). Therefore, the high
spatial and temporal resolution of aerial photographs and
their low cost make them an effective tool for studying
changes in forests (Verheyden et al. 2002; Honey-Rose
´s
et al. 2009). Of the core zone of the MBBR, 81.0% has
some type of vegetation; among the predominant commu-
nities, 42.8% is fir, 29.3% pine and 15.8% oak (Lo
´pez-
Garcı
´a and Vega 2010). The high precision obtained in this
part of the process is due to the use of high-resolution aerial
photographs and to the experience of the photo interpreters
(Lo
´pez-Garcı
´a et al. 2016). Therefore, the objective of this
current study was to determine the biennial change
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
processes of a temperate forest in the MBBR using high-
resolution digital aerial photographs to determine the
changes in disturbance or forest recovery.
Materials and methods
Study area
The MBBR is in central Mexico, divided between the
states of Me
´xico and Michoaca
´n (19°5900300 to 19°2101500
N and 100°0702600 to 100°2000700 W) (Fig. 1), and has an
area of 56 259 ha divided into two buffer zones (42
704 ha) and three core zones (13 555 ha). [It lies on
Neovolcanic Andesites consisting wholly or partly of
pyroclastic products, which have developed into volcanic
soils.] The climate is humid temperate with summer rains
(Lo
´pez-Garcı
´a and Alca
´ntara-Ayala 2012). The core zone
has an elevation from 2200 m at the western end to 3600 m
in the Sierra Campanario. This increasing altitude is
associated with a vegetation gradient from broad-leaved
forests dominated by oaks (Quercus leiophylla A. DC, Q.
Fig. 1 Location of the study area in the states of Me
´xico and Michoaca
´n (Mexico)
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
laurina Humb et Bonpl., Q. obtusata Bonpl. and Q. rugosa
Ne
´e), through communities of pine (Pinus pseudostrobus
Lindl., P. montezumae Lamb., P. oocarpa Shiede. and P.
michoacana Mart.) and, finally, to fir forests (Abies reli-
giosa Kunth Schltdl. y Cham.) at altitudes of 2900–3600 m
(Lo
´pez-Garcı
´a and Vega 2010).
Collection of data
Between 1999 and 2013, 14 aerial photographic surveys
were carried out in the MBBR. The photographs were in
colour with increasing spatial resolution (1.1 m for 1999,
0.85 m in 2001, 0.6 in 2003, 0.4 m in 2005-2007 and 0.3 m
as of 2009) as a result of technological improvements of
cameras and lenses (Lo
´pez-Garcı
´a et al. 2016).
Orthorectified mosaics were generated from digital
aerial photographs for 1999 and 2013 with a resolution of
1 m/pixel, with support from orthophotographs at 2 m/
pixel provided by the National Institute of Statistics,
Geography and Informatics (INEGI). A digital terrain
model with contours every 20 m, UTM and a WGS84
datum projection with the support of 30 control points on a
smooth ground surface, and 40 on a rough surface were
used. The mosaics were constructed with ERDAS with a
mean square error between 7 and 5 m respectively.
The same flight plan was used in every survey, over-
flying the same 11 lines and applying the same conven-
tional techniques as in photographic imagery for generation
and stereo viewing, thereby guaranteeing an overlap of
30% laterally and 60% longitudinally (Lo
´pez-Garcı
´a
2009). The digital aerial photographs were interpreted on a
1:10,000 scale with a mirror stereoscope, considering five
levels of tree cover density (Lo
´pez-Garcı
´a2011), and a
minimum mappable area of 650 m
2
for deforested areas
and land uses, and 5000 m
2
for forest cover.
The vegetation cover was defined by the values of forest
cover and land use from photo interpretation based on
shape, size, texture, tone, colour, shade of trees and the
separation between the tops of the forest cover observed in
the stereoscopic model (Avery 1969). Areas with [75%
canopy were classified as closed forest, 51–75% as semi-
enclosed, 26–50% as semi-open, 10–25% as open (Lo
´pez-
Garcı
´a2011) and \10% as deforested (FAO 2010).
Density of forest cover
This is the basis to determine the processes involved in
cover change and in forest dynamics. These processes are
related in such a way that the definition of forest establishes
three fundamental aspects: forests have an area of more
than 5000 m
2
, a canopy cover of more than 10%, and trees
greater than 5 m in height (FAO 2012). Processes con-
tributing to change, such as forest degradation or forest
improvement, occur within forests with [10% coverage.
Reforestation occurs when the forest cover is temporarily
\10%. Deforestation and afforestation represent transfers
between forests and other kinds of land use (FAO 2000).
The definitions of the different processes of change (forest
degradation, deforestation, forest improvement, reforesta-
tion and afforestation) are based on FAO (2000). A change
in land use refers to the transfer of forest to another land
use such as arable farming, livestock grazing or urban
cover.
Biennial interpretation
Changes between biennial series were determined by
comparative photo interpretation of pairs of digital aerial
photographs, 1999–2001, 2001–2003, 2003–2005,
2005–2007 and 2007–2009, to determine changes in the
canopy. The polygons were transferred by radial triangu-
lation in digital format at an average scale of 1:5000
(Lo
´pez-Garcı
´a2009). In this way, cover changes were
obtained as a biennial series. It should be noted that some
areas of change overlap a degradation process where the
forest cover was reduced biennially.
General interpretation
Based on the different densities of forest cover (closed,
semi-closed, semi-open, open and deforested) and types of
land use, the digital aerial photographs of 2013 were
evaluated. The photo interpretation was transposed by
radial triangulation onto the ortho mosaic of 2013 made
with the same photographs using the features common to
both photographs and the ortho mosaic. Once the polygon
map of 2013 was finished, a copy was taken and the ortho
mosaic of 1999 was superimposed to modify the polygons
of the zones of change between 1999 and 2013, and the
1999 forest cover density map was obtained.
Field verification
Photo interpretation was carried out between March and
April of each year; subsequently, field verification in 2001,
2003, 2005, 2007, 2009, 2011 and 2013, during which all
changes were verified, except for the year 2007. Lack of
verification in 2007 in the field was due to the reticence of
the community of Crescencio Morales due illegal logging
that affected 731 ha of dense forest. Therefore, verification
was made based on digital aerial photographs. Verification
was carried out with the support of the representatives of
the communities as local guides; non-governamental
organizations: World Wild Life Fund (WWF), and Mexi-
can Fund for the Conservation of Nature (Fondo Mexicano
para la Conservacio
´n de la Naturaleza, FMCN);
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
government institutions; Federal Attorney for Environ-
mental Protection (Procuraduria Federal de Proteccio
´nal
Ambiente, PROFEPA), Ministry of Environment and
Natural Resources (Secretarı
´a de Medio Ambiente y
Recursos Naturales, SEMARNAT), National Commission
of Natural Protected Areas (Comisio
´n Nacional de A
´reas
Naturales Protegidas, CONANP) and National Forestry
Commission (Comisio
´n Nacional Forestal, CONAFOR).
For the overall change analysis, verification routes were
carried out for the 1999 coverage after the photo inter-
pretation of the vegetation and the coverage density
between March and June 2000. For the photo interpretation
and coverage density of 2013, verification was between
March 2013 and November 2015, including photo inter-
pretation and classification of the vegetation types. In 73
1000 m
2
circular sample plots, the type of tree cover,
diameters and heights were determined.
Change matrices
In the biennial analyses, the negative changes represented
by the combination of anthropogenic effects (illegal log-
ging) and changes due to natural effects (hydrometeoro-
logical events, pests and diseases) were evaluated. The
analysis of changes between 1999 and 2013 detected by
combining forest covers of 1999 and 2013, resulted in a
matrix showing negative changes, (above the diagonal),
and positive changes (below the diagonal). The two
matrices combined gave a more complex matrix that
allowed for the differentiation of several processes within
the forest dynamics.
An error matrix was elaborated from the vegetation
map, field trips and 1000 m
2
circular samples. Another
error matrix was also made for the map of density of forest
cover, field trips and 1000 m
2
circular sampling. to deter-
mine the accuracy of the producer and of the user, errors of
omission and commission of each of the maps. From the
same error matrix, the overall accuracy was calculated
(data in the diagonal divided by the total data).
Results
Biennial changes
Based on the digital aerial photographs, the biennial anal-
ysis allowed for the evaluation of deterioration due to
forest degradation and to deforestation (Fig. 2). In turn,
these were divided into anthropogenic and natural effects,
the former due to illegal logging and represented by areas
with obvious changes such as branch and leaf residues,
bare soil, soil removal and recently-built roads (Fig. 3).
Changes of natural origin due to extreme
hydrometeorological events, such as debris flows, mud
flows and windfall, are easily distinguished in aerial pho-
tographs. To a lesser extent, the effects of drought can be
discerned as patches with a change in the tonality of the
forest canopy.
The biennially accumulated area changes between 1999
and 2013 amounted to 2274.2 ha but 343.4 ha underwent
processes of repeated change (change over two periods on
the same site). In the first period, 1999–2001, 406.1 ha
were deforested, followed by a decrease to 288.5 ha during
2001–2003, an increase to 463.6 ha during 2003–2005, and
to 731.1 ha during 2005–2007. The change recorded in this
area was markedly reduced to 237.2 ha in the period
2007-2009 (Table 1). From this period onwards there was a
considerable decrease in illegal logging to only 5.1 ha in
2009–2011. The other change in area, (114.9 ha), was due
to climatic effects. In the final period (2011–2013), the
changes affected 27.9 ha, mainly caused by pests, diseases
and drought in the upper parts of the mountains and only
5 ha due to illegal logging. Most of the biennial changes
occurred in the central part of the core zone (Fig. 2).
Changes in forest cover between 1999 and 2013
The changes over this period were derived from the
changes matrix (Table 2). There were six change processes
identified (Fig. 4; Table 3). The processes of disturbance
were forest degradation, deforestation and land use change.
The recovery processes were forest improvement, refor-
estation and afforestation.
The total change between 1999 and 2013 in forest cover
was 4902.1 ha (36.2%). The loss of forest cover was
21.5%, mainly due to deforestation and reduction of tree
density. The recovery processes involved 14.7% of the
total area, mainly due to an increase in the density of trees
and to afforestation. The unchanged area was 8653.1 ha
(63.8%; Tables 2and 3; Fig. 4).
Changes from one category to the next (closed to semi-
closed; semi-closed to semi-open) represent 59.4% of the
forest loss; the remaining was due to more severe changes
in forest cover of three or four category levels (closed to
open or closed to deforested).
Between 1999 and 2013, closed canopy recovered with
an annual rate of 0.2% and, in contrast, semi-closed and
semi-open hedges were transformed at a rate of -3.6%
and -4.2%, respectively. The open coverage had an
annual exchange rate of 2.8% and the deforested category a
rate of 4.9% (Table 2). The best preserved and largely
unchanged woodlands were located in the north of the
Sierra Chincua and on the southern slopes of Cerro Pelo
´n
(Fig. 1).
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
Fig. 2 Biennial forest cover
changes in the core zone of the
Monarch Butterfly Biosphere
Reserve between 1999 and 2013
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
Change matrix (biennial and general)
The combined biennial and general change matrix revealed
gradual changes that were not detectable in the biennial
analyses alone. The combination of the processes of change
involved in the forest dynamics was included. This matrix
facilitated understanding of part of the forest dynamics
because the biennial changes or the gradual recoveries
showed losses or gains, revealing the different states of the
forest over time, as well as selective logging. The matrix of
19 categories reflects different processes within the forest
cover changes (Fig. 5) and shows an unchanged area of
8535.6 ha (63.0%) in the core zone. Changes in the forest
cover of 5019.6 ha were related to nine common change
processes occurring in both matrices involving forest losses
and recoveries (Table 4).
Some changes were not detected in the biennial analysis
(column ‘‘no change’’, Table 4), but were identified when
considering the whole period which revealed gradual
changes to 3088.9 ha, including cover loss, deforestation,
degradation and land use changes at different levels of
intensity (Tables 2and 4). In addition, recovery was evi-
dent in 1877.4 ha due to forest improvement processes,
reforestation and afforestation (Fig. 5and Table 4). The
matrix also shows biennial changes that did not appear in
the analysis of the entire period (row ‘‘no change’’,
Table 4) due to negative changes in the initial two-year
period.
The change processes were grouped into five types: (1)
large-scale logging (1570.9 ha); (2) small-scale logging
(1211.5 ha); (3) natural changes due to climatic effects
(133.5 ha), including landscape recovery processes that
were subsequently influenced by climatic effects; (4)
Fig. 3 Forest cover changes related to deforestation and degradation caused by large-scale and small-scale logging in the Ejido of Crescencio
Morales, Michoaca
´n
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
changes in recovery (1986.3 ha), including landscapes that
suffered deforestation and degradation in the initial bien-
nial periods and subsequently experienced retrieval; and,
(5) changes in loss-recovery (117.5 ha), which apparently
had not changed at the end of the period but decreased the
quality and composition of the forest (Table 5and Fig. 6).
Changes in vegetation
The changes were 4902.1 ha as follows: pine forests (Pinus
spp.), 53.5% (losses 30.3%, recoveries 23.2%); fir forests
(Abies religiosa), 29.6% (losses 19.6%, recoveries 10.0%);
oak forests (Quercus spp.), 16.0% (losses 6.2%, recoveries
6.6%); and other forest types, cedar (Cupressus spp.), alder
Table 1 Biennial forest cover changes 1999–2013 in the core zone of the Monarch Butterfly Biosphere Reserve (ha)
Change in forest
density
Biennial period Total
accumulated
Total
1999–2013
1999–2001 2001–2003 2003–2005 2005–2007 2007–2009 2009–2011 2011–2013
Closed—semi-
closed
66.2 36.0 62.0 33.4 17.7 5.4 7.6 228.2 296.4
Closed—semi-
open
25.3 28.1 69.4 71.3 25.3 8.0 2.3 229.6 233.7
Closed—open 3.5 1.5 31.7 107.3 22.6 6.1 1.8 174.5 220.4
Semi-closed—
semi-open
34.1 20.5 39.1 44.7 14.4 19.5 6.4 178.6 250.9
Semi-closed—
open
14.6 29.4 22.7 52.8 10.9 12.3 2.3 145.0 163.3
Semi-open—
open
54.2 20.7 22.6 43.8 21.1 6.5 5.4 174.3 253.3
Subtotal
degradation
197.9 136.2 247.3 353.3 112.0 57.8 25.8 1130.2 1418.0
Closed—
deforested
56.6 23.4 42.8 152.7 32.4 20.9 0.00 328.9 598.4
Semi-closed—
deforested
17.9 17.8 19.2 56.9 16.8 14.2 0.5 143.3 253.5
Semi-open—
Deforested
27.1 62.5 61.5 93.2 40.0 23.0 0.9 308.2 400.9
Open—
deforested
106.7 48.6 92.8 75.0 35.6 4.3 0.7 363.7 240.5
Subtotal
deforested
208.3 152.3 216.4 377.8 124.8 62.4 2.1 1144.0 1493.2
Total 406.2 288.5 463.7 731.1 236.8 120.2 27.9 2274.2 2911.2
Table 2 Matrix of overall
forest cover changes between
1999 and 2013 in the core zone
of the Monarch Butterfly
Biosphere Reserve
Categories (ha) 2013 Total 1999
Closed Semi-closed Semi-open Open Deforested No forest
1999 Closed 6407.6 296.4 233.7 220.4 598.4 0.0 7756.4
Semi-closed 577.8 616.7 250.9 163.3 253.5 0.0 1862.2
Semi-open 563.6 107.4 310.7 253.3 400.9 0.5 1636.5
Open 101.0 57.8 34.3 268.8 240.5 0.0 702.2
Deforested 273.4 37.7 59.6 131.5 550.1 0.3 1052.6
No forest 30.3 5.8 1.3 1.4 7.5 499.1 545.3
Total 2013 7953.7 1121.8 890.4 1038.7 2050.8 499.9 13555.2
The diagonal shows the categories that remained unchanged. Above the diagonal are disturbance processes
and below are recovery processes
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
Fig. 4 Forest cover changes
defined in the core zone of the
Monarch Butterfly Biosphere
Reserve between 1999 and 2013
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
(Alnus spp.) and juniper (Juniperus monticola), 0.9%
(losses 0.2%, recovery 0.7%) (Table 6).
Regarding forest cover, the pine forests were disturbed
over an area of 1487.2 ha, of which 471.7 ha were closed
and semi-closed forests and were deforested, and 516.4 ha
suffered degradation. In the fir forests, 958.9 ha were dis-
turbed, of which 774.8 ha were closed and semi-closed
forests which were affected, including deforestation
(349.1 ha) and forest degradation (426.3 ha). Regarding
the oak forests, 458.3 ha were disturbed, with a trend
Table 3 Overall forest cover
changes between 1999 and 2013
in the core zone of the Monarch
Butterfly Biosphere Reserve
General processes Surface (ha) Relative (%) Absolute (%)
Disturbance Deforestation 1493.2 30.5 11.0
Forest degradation 1418.0 28.9 10.5
Land use change 0.8 0.0 0.0
Subtotal losses 2912.0 59.4 21.5
Recovery Forest improvement 1441.8 29.4 10.6
Reforestation 502.2 10.2 3.7
Afforestation 46.2 0.9 0.3
Subtotal gains 1990.2 40.7 14.7
Total 4902.1 100.0 36.2
Fig. 5 Revegetation processes and densification due to reforestation in different years in the Ejido Hervidero y Plancha, Michoaca
´n, Me
´xico
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
towards increasing degradation and deforestation. Finally,
7.5 ha of juniper forest were disturbed.
Error analysis
The estimation of the density of forest cover had an overall
accuracy of 95.65% and a Kappa coefficient of 93.3%.
Omission and commission errors occurred for the semi-
closed coverage (5.6%) and the semi-open and open cov-
erages (14.29%). In the closed and deforested categories,
there were no errors of omission or commission. Photoin-
terpretation identified the vegetation types in the core zone
with an overall accuracy of 90.2% and a Kappa coefficient
of 84.8%, with an omission error of 6.0% and a commis-
sion error of 7.8% for the fir forest. The pine, oak and cedar
forests had omission errors of 12.5, 20 and 25%,
respectively, and commission errors of 17.6, 0 and 25%,
respectively.
Discussion
Studies in the MBBR using digital high-resolution aerial
photographs have allowed calibration of this method and
reliable and accurate determination of forest cover changes
(Brower et al. 2002;Lo
´pez-Garcı
´a2009;Lo
´pez-Garcı
´a
et al. 2014,2016; Manzo-Delgado et al. 2014; Vidal et al.
2014). The method uses a sequence of steps: stereoscopic
rendering of comparative photointerpretation, interpreta-
tion in digital format, field evaluation and digital restitution
on orthorectified mosaics (Lo
´pez-Garcı
´a et al. 2016). In
this study, photointerpretation of the cover density had very
high overall accuracy and Kappa coefficient. This is
Table 4 Matrix of biennial changes and overall forest cover changes between 1999 and 2013 in the core zone of the Monarch Butterfly
Biosphere Reserve
General
change (1999
and 2013)
Biennial changes (1999-2013) General
total
(ha)
Cover change
processes
No change Deforestation by
logging
Forest degradation by
logging
Deforestation by
natural
phenomena
Forest degradation by
natural phenomena
No change No change
(8535.6 ha)
Deforestation—
reforestation—No
change (42.2 ha)
Forest degradation
forest
improvement—No
change (68.9 ha)
Reforestation—
deforestation—
no change
(0.9 ha)
Forest improvement—
forest degradation—
no change (5.4 ha)
8653.1
Deforestation
by logging
Gradual
deforestation
(433.3 ha)
Deforestation
(861.4 ha)
Forest degradation—
Deforestation
(139.1 ha)
Deforestation by
natural
phenomena
(59.4 ha)
1493.2
Forest
degradation
by logging
Gradual Forest
degradation
(777.4 ha)
Deforestation—
reforestation—forest
degradation
(54.7 ha)
Forest degradation
(515.6 ha)
Forest degradation by
natural phenomena
(70.2 ha)
1418.0
Land use
change
Gradual
deforestation—
land change use
(0.8 ha)
0.8
Forest
improvement
Gradual Forest
improvement
(1332.8 ha)
Deforestation—
Reforestation—
forest improvement
(84.01 ha)
Forest degradation—
forest improvement
(24.8 ha)
1441.8
Reforestation Reforestation
(498.3 ha)
Reforestation—forest
degradation by
natural phenomena
(3.9 ha)
502.2
Afforestation Afforestation
(46.2 ha)
46.2
Total biennial
(ha)
11624.5 1042.4 748.5 60.3 79.5 13555.2
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
because the interpretation of the aerial photographs in color
at a 1: 10 000 scale used to separate the closed category is
easily distinguishable by the continuous forest canopy
supported by the elements of photo interpretation. The
deforested category is the clear tone and the absence of
shadows made its delimitation easy. Nevertheless, the
semi-closed, semi-open and open categories were confused
with other categories as they resulted from a qualitative
evaluation; therefore, errors of omission and commission,
as well as the similarity between these values (for example,
semi-closed versus semi-open or semi-open versus semi-
closed), have to be considered.
The photo interpretation of vegetation types was less
successful, with a general accuracy of 78.7%. However, for
the fir forest, which is most widely distributed with 35.0%
occurring in the core zone, the omission and commission
errors were low (2.3% and 10.4%, respectively).
In this study, photo interpretation has shown its value as
a cartographic product to identify vegetation types and
although it is time-consuming, it has precision and relia-
bility. Accuracy increases with digital aerial colour pho-
tographs and scales higher than or equal to 1:10 000, as
well as with the use of stereoscopy to interpret species
composition and forest cover density (Lo
´pez-Garcı
´a et al.
2016).
The biennial analysis only determined changes due to
disturbance, whereas the general analysis for the 14-year
period detected changes due to both disturbance and to
forest recovery. However, the combination of both analyses
allowed for the detection of gradual processes of both
disturbance and recovery which complemented the analysis
of the changes.
For small-scale logging and gradual negative changes,
this method illustrates the importance of their separation
because the magnitude of these changes could alert us to
the potential for more drastic changes. The analysis of the
overall forest cover changes between 1999 and 2013
showed a constant closed cover, but some areas reflected a
decrease in the quality of the forest due to recent
reforestation.
This extension of the study has produced more detailed
knowledge of the changes in the landscape and will aid the
management of the MBBR and priority areas for restora-
tion and reforestation.
The biennial analysis led to the identification of rapid
change processes related to forest degradation, deforesta-
tion and land use change, distinguishing four levels of
disturbance, depending on the resulting forest cover. The
logging that has occurred in this region, combined with
intense hydrometeorological events such as those of
February 2010, that caused mudflows, has had disastrous
effects (Alca
´ntara-Ayala et al. 2012). Likewise, the lack of
forest management has allowed the forest to be attacked by
plagues (Scolytus mundus Wood y Dendroctonus mexi-
canus Hopkins) and diseases (Gardun
˜o2011).
Since timber harvesting is not allowed in the core zone
of the Reserve, the decrease in tree density is due to
clandestine logging, which leads to forest degradation and,
in severe cases, to deforestation. Comparison of the bien-
nial matrix with the 14-year matrix detected changes due to
small-scale logging that were not evident in the individual
matrices.
The total loss of forest was 6.8% in the MBBR core
zone, representing an annual loss of 0.5%. Large-scale
Table 5 Causes related to forest cover changes in the core zone of the Monarch Butterfly Biosphere Reserve
Kind of change Processes of change Surface (ha) Relative (%) Absolute (%)
Large-scale logging Forest degradation 1000.5 19.9 7.4
Deforestation 570.4 11.4 4.2
Small-scale logging Gradual forest degradation 433.3 8.6 3.2
Gradual deforestation 777.4 15.5 5.7
Land use change 0.8 0.0 0.0
Climate-related Forest degradation by natural phenomena 59.4 1.2 0.4
Reforestation-Forest degradation by natural phenomena 70.2 1.4 0.5
Deforestation by natural phenomena 3.9 0.1 0.0
Recovery Forest improvement 1332.9 26.5 9.8
Reforestation 498.3 9.9 3.7
Afforestation 46.2 0.2 0.3
Deforestation-Reforestation 84.1 1.7 0.6
Forest degradation-Forest improvement 24.9 0.5 0.2
Total 4902.1 100.0 36.2
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
Fig. 6 a Forest degradation; bdeforestation; cdeforestation—Forest degradation; dmud flow; eforest improvement; freforestation;
greforestation—Forest improvement; hafforestation
Table 6 Overall forest recovery and forest losses during the period 1999–2013 according to the dominant vegetation type, in the core zone of the
Monarch Butterfly Biosphere Reserve
Dominant forest type
(ha)
Deforestation Forest
degradation
Subtotal
Losses
Forest
improvement
Reforestation Afforestation Subtotal
Gains
Total
Abies
a
488.6 470.3 958.9 417.6 72.1 2.5 492.1 1451.0
Pinus
a
844.4 642.9 1487.3 799.6 319.8 18.1 1137.5 2624.7
Quercus
a
154.1 304.2 458.3 213.2 99.5 13.0 325.7 784.0
Others
b
6.9 0.6 7.5 11.4 10.9 12.6 34.9 42.4
Total 1494.0 1418.0 2912.0 1441.8 502.2 46.2 1990.2 4902.1
a
Dominant genus,
b
Includes Cupresus, Juniperus, Alnus as the dominant genus
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
logging is the main reason for forest cover changes (Mer-
ino-Pe
´rez and Herna
´ndez-Apolinar 2004). Logging causes
deforestation and degradation of well-conserved forests,
particularly in local communities not participating in the
Payment for Environmental Services (PES) programme
(Vidal et al. 2014). The selective logging practiced by
farming communities and characterised by regular and
constant logging for domestic use is regarded as small-
scale logging, but it has led to forest losses of 93.2 ha/year.
The origin of these logging activities seems to be due to
the proposal to increase the Reserve area and has prompted
the communities to harvest the forest. Once the MBBR had
been established, not all the agricultural communities
participated in the PES program. Therefore, logging
activities continued in communities that were not included
in this program or arose through disputes between com-
munities, reaching a peak in 2005–2007, during which
731 ha of conserved closed forests were felled. Defor-
estation in one of the best preserved areas in the Reserve
(Lomas de Aparicio) has caused the disappearance of
monarch butterfly colonies (Lo
´pez-Garcı
´a2011). After
intense collaboration with communities reluctant to incor-
porate land into the PES program, there was a decrease in
illegal logging which reached its minimum historical level
(5 ha) in the years 2009–2011 (Vidal et al. 2014). These
changes deserve particular attention because the PES
scheme implemented by the Monarch Fund makes an
annual payment to communities based on the absence of
logging and, therefore, on forest conservation with cover-
age greater than 50%. But gradual changes are not detected
until several years have passed and the scheme must be
evaluated every 5 years. This assessment contrasts with the
forest loss due to small-scale logging within the core zone
of the MBBR, which should prompt a change in the PES in
this Reserve.
The combination of the change matrices identified 19
processes involved in the forest cover dynamic, high-
lighting the recovery processes that are the result of PES
programs, and reforestation and forest conservation pro-
grams conducted by forest owners. These programs, the
result of long negotiations, have raised awareness and
improved the situation of the 34 agrarian communities that
signed the contract with the Monarch Fund. Therefore, the
legal protection of the core zone and the financial incen-
tives granted by the Monarch Fund have supported the
conservation of the hibernation forests of the monarch
butterfly (Honey-Rose
´s et al. 2009). Between 2001 and
2009, 1721 ha of forests were affected (Vidal et al. 2014);
in response, in 2009, the reforestation programs were
intensified in areas affected by illegal logging, and 29
forest nurseries were created with a production capacity of
6,508,453 seedlings per year (Venegas et al. 2011).
Surveillance committees were organised in all the agrarian
communities and annual reforestation programs were
established. Between 2009 and 2010, 3769.5 ha were
reforested in the 11 agrarian communities most affected by
logging.
The designation of the area as the MBBR, accompanied
by monitoring programs and the PES, has been followed by
forest conservation, leading to different levels of recovery
and an increase in forest cover (forest improvement). The
management of deforested areas has led to the recovery of
forest land (reforestation). Finally, non-forested areas,
(mainly agricultural land), have been transformed by
reforestation, mainly with pines and to a lesser extent with
cedars.
The pine forests were those most affected by logging
due to the commercial value of their timber, but they also
had the highest recovery rate compared to regular refor-
estation. The fir forests recovered only half the area lost
during the study period, mainly by forest improvement, and
the oak forests showed the greatest recovery due to natural
processes. The Juniperus monticola forests face a serious
conservation threat and are subject to special protection
(SEMARNAT 2010). These coniferous shrubs are located
at altitudes above 2900 m, forming pure stands within
Abies religiosa forests and are associated with natural
grasslands. The area of this forest has declined due to
extraction of firewood for domestic use.
On balance, comparing the area of forest lost with the
area recovered, or at least tending towards positive
recovery in the period 1999–2013, overall loss in the core
zone of the MBBR, 6.8%, represents an annual loss of
0.5%; this must be countered by continued diligence and a
willingness to adapt the PES scheme in response to more
appropriate monitoring.
Conclusion
This study presents forest cover changes in the core zone of
the Monarch Butterfly Biosphere Reserve in terms of
sequential processes before and after the designation of the
Reserve and the implementation of PES programs between
conservation institutions and local communities. Even
small-scale logging should be considered as a serious
problem that can reduce the quality of the forests. The
disturbance processes must be reversed and recovery pro-
cesses enhanced to guarantee the permanence of the
migratory process of the monarch butterfly. It is also nec-
essary to protect the forests of Juniperus monticola with
special protection status due to their ecological function.
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123
However, there is a promising trend towards forest recov-
ery, since local communities have become more involved
in surveillance and reforestation programs within the core
zone of the MBBR.
The establishment of a Protected Area must guarantee
the conservation of natural resources but it must also be
socially accepted, technically possible and ecologically
sustainable.
References
Alca
´ntara-Ayala I, Lo
´pez-Garcı
´a J, Garnica R (2012) On the landslide
event in 2010 in the Monarch Butterfly Biosphere Reserve,
Angangueo, Michoaca
´n, Mexico. Landslides 9:263–273
Anderson-Teixeira KJ, DeLucia EH (2011) The greenhouse gas value
of ecosystems. Glob Change Biol 17:425–438
Angelsen A, Wertz-Kanounnikoff S (2008) What are the key design
issues for REDD and the criteria for assessing options? In:
Angelsen A (ed.) Moving ahead with REDD: Issues, options and
implications, pp 11–21. Bogor, Indonesia: Center for Interna-
tional Forestry Research (CIFOR)
Avery TE (1969) Forester’s guide to aerial photo interpretation.
Agriculture Handbook No. 308, p 40. U.S. Department of
Agriculture. Washington, D.C.: Forest Service. U.S. Government
Printing Office
Brower LP, Castilleja G, Peralta A, Lo
´pez-Garcı
´a J, Bojorquez-Tapia
L, Diaz S, Melgarejo D, Missrie M (2002) Quantitative changes
in forest quality in a principal overwintering area of the monarch
butterfly in Mexico, 1971–1999. Conserv Biol 6(2):346–359
Brower L, Taylor O, Williams E, Slayback D, Zubieta R, Ramı
´rez M
(2012) Decline of monarch butterflies overwintering in Me
´xico:
is the migratory phenomenon at risk? Insect Conserv Divers
5:95–100
Champo-Jime
´nez O, Valderrama-Landeros L, Espan
˜a-Boquera M
(2012) Pe
´rdida de cobertura forestal en la Reserva de la Bio
´sfera
Mariposa Monarca, Michoaca
´n, Me
´xico. (2006–2010). Revista
Chapingo. Serie Ciencias Forestales y del Ambiente
18(2):143–157
Dahdouh-Guebas F, Verheyden A, De Genst W, Hettiarachchi S,
Koedam N (2000) Four decades vegetation dynamics in Sri
Lankan mangroves as detected from sequential aerial photogra-
phy: a case study in Galle. Bull Mar Sci 67(2):741–759
FAO (2000) On definitions of forest and forest change. Forest
resources assessment. Working Paper 33, 14p. Rome, Italia
FAO (2010) Global forest resources assessment 2010. Main report.
FAO Forestry Paper No. 163, 378p. Rome, Italy
FAO (2012) Terms and definitions. Forest resources assessment
Working Paper 180,31p. Rome, Italy
FAO (2015) Global Forest Resources Assessment 2015. Desk
Reference. (Main report.) FAO. Rome, Italy, 253p
Fensham R, Fairfax R (2002) Aerial photography for assessing
vegetation change: a review of applications and the relevance of
findings for Australian vegetation history. Aust J Bot
50:415–429
Figueroa F, Sa
´nchez-Cordero V, Illoldi-Rangel P, Linaje M (2011)
Evaluation of protected area effectiveness for preventing land
use and land cover changes in Mexico. Is an index good enough?
Revista Mexicana de Biodiversidad 82:951–963
Gardun
˜o BN (2011) Diagno
´stico fitosanitario forestal. Reserva de la
Biosfera Mariposa Monarca en el Estado de Me
´xico, Me
´xico, p
p106
Honey-Rose
´s J (2009) Illegal logging in common property forests.
Soc Nat Res 22:916–930
Honey-Rose
´sJ,Lo
´pez-Garcı
´a J, Rendo
´n-Salinas E, Peralta-Higuera
A, Galindo-Leal C (2009) To pay or not to pay? Monitoring
performance and enforcing conditionality when paying for forest
conservation in Mexico. Environ Conserv 36:120–128 http://
www.ciesin.columbia.edu/docs/002-471/002-471.html. Acces-
sed 09 May 2015
Lo
´pez-Garcı
´a J (2009) Ana
´lisis de cambio en la cobertura forestal en
la zona nu
´cleo de la Reserva de la Biosfera Mariposa Monarca:
2003-2006. In: Conceptos y aplicaciones de la geoma
´tica en
Me
´xico. Me
´xico: Instituto de Geografı
´a, Universidad Nacional
Auto
´noma de Me
´xico, [Aguirre R (ed.)] pp 195–210
Lo
´pez-Garcı
´a J (2011) Deforestation and forest degradation in the
Monarch Butterfly Biosphere Reserve, Mexico, 2003–2009.
J Maps 7(1):626–633
Lo
´pez-Garcı
´a J, Alca
´ntara-Ayala I (2012) Land-use change and
hillslope instability in the Monarch Butterfly Biosphere Reserve,
Central Mexico. Land Degrad Dev 23:384–397
Lo
´pez-Garcı
´a J, Vega A (2010) Vegetation and land use 2009:
Monarch Butterfly Biosphere Reserve,Mexico.
J Maps 6(1):665–673
Lo
´pez-Garcı
´a J, Manzo-Delgado L, Alca
´ntara-Ayala I (2014) Rural
aquaculture as a sustainable alternative for forest conservation in
the Monarch Butterfly Biosphere Reserve, Mexico. J Enviro
Manag 138:43–54
Lo
´pez-Garcı
´a J, Prado Molina J, Manzo-Delgado LL, Peralta Higuera
A (2016) Monitoring changes of forest canopy density in a
temperate forest using high-resolution. Boletı
´n del Instituto de
Geografı
´a, Investigaciones Geogra
´ficas 90:59–74
Lund HG (2014) What is a forest? Definitions do make a difference:
an example from Turkey. Avrasya Terim Dergisi 2(1):1–8
Manzo-Delgado L, Lo
´pez-Garcı
´a J, Alca
´ntara-Ayala I (2014) Role of
forest conservation in lessening land degradation in a temperate
region: the Monarch Butterfly Biosphere Reserve, Mexico.
J Environ Manag 138:55–66
Merino-Pe
´rez L, Herna
´ndez-Apolinar M (2004) Destruccio
´nde
instituciones comunitarias y deterioro de los bosques de la
Reserva de la Biosfera Mariposa Monarca, Michoaca
´n, Me
´xico.
Revista Mexicana de Sociologı
´a 66:261–309
Miller AD, Dietze MC, DeLucia EH, Anderson-Teixeira KJ (2016)
Alteration of forest succession and carbon cycling under
elevated CO2. Glob Change Biol 22:351–363
Navarrete JL, Ramı
´rez M, Pe
´rez-Salicrup D (2012) Logging within
protected areas: spatial evaluation of the monarch butterfly
biosphere reserve, Mexico. For Ecol Manag 262:646–654
Pan Y, Richard A, Birdsey RA, Fang J, Houghton R, Kauppi PE, Kurz
WA, Phillips OL, Shvidenko A, Lewis SL, Canadel JGI, Ciais P,
Jackson RB, Pacala SW, McGuire AD, Piao S, Rautiainen A,
Sitch S, Hayes D (2011) A large and persistent carbon sink in the
world’s forests. Science 333:988–993
Ramı
´rez MI, Gime
´nez A, Luna L (2003) Effects of human activities
on monarch butterfly habitat in protected mountain forest,
Mexico. For Chron 79:242–246
SEMARNAT (2010) Norma Oficial Mexicana NOM-059-SEMAR-
NAT-2010, Proteccio
´n ambiental– Especies nativas de Me
´xico
de flora y fauna silvestres– Categorı
´as de riesgo y especifica-
ciones para su inclusio
´n, exclusio
´n o cambio– Lista de especies
en riesgo. Diario Oficial de la Federacio
´n 30 diciembre, 2010.
Me
´xico, 77p
Venegas PY, Rodrı
´guez S, Lo
´pez DT (2011) Ana
´lisis base para el
disen
˜o de la estrategia de reforestacio
´n de la Reserva de la
Biosfera Mariposa Monarca. http://monarchbuterflyfund.org/
sites/default/files/Analisis_Reforestacion_RBMM_MBF_2011.
pdf. Accessed 19 June 2015
Disturbance and forest recovery in the Monarch Butterfly Biosphere Reserve, Mexico
123
Verheyden A, Dahdouh-Guebas F, Thomaes W, De Genst S,
Hettiarachchi S, Koedam N (2002) High Resolution vegetation
data for mangrove research as obtained from aerial photography.
Environ Dev Sustain 4(2):113–133
Vidal O, Lo
´pez-Garcı
´a J, Rendo
´n-Salinas E (2014) Trends in
deforestation and forest degradation after a decade of monitoring
in the Monarch Butterfly Biosphere Reserve in Mexico. Conserv
Biol 28:177–186
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
J. Lo
´pez-Garcı
´a, R. M. Navarro-Cerrillo
123