Forest depletion gradient along the Amazon floodplain

Ecological Indicators 98 (2019) 409–419 Contents lists available at ScienceDirect Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind Forest depletion gradient along the Amazon floodplain ⁎ T Vivian Renó , Evlyn Novo Remote Sensing Division, Earth Observation Coordination (OBT), Brazilian Institute for Space Research, Av. Astronautas, 1.758 - Jardim da Granja, São José dos Campos, São Paulo, Brazil A R T I C LE I N FO A B S T R A C T Keywords: Amazon floodplain Deforestation Forest fragmentation Landscape metrics Remote sensing Landsat time-series This article analyzes the process of forest cover depletion over the last 40 years at three landscapes distributed along the Amazon floodplain. To this end, we created multi-temporal forest cover maps based on time series of Landsat images, and then analyzed the forest cover dynamics through landscape metrics. Based on landscape analyzes and bibliographic information, we assessed the degree of forest depletion of each landscape and made inferences regarding the main drives of forest changes and their impacts on ecosystem integrity. Results show the existence of an east-to-west gradient of forest depletion that varies in time and space along the floodplain of the Solimões/Amazonas River, and provides evidence that it is a response to the history of human occupation and public policies. The most degraded landscapes are located on the eastern region, where forest depletion degree indicates substantial damage to biodiversity and ecosystem services. The study increases the scarce knowledge about the dynamics of the floodplain forest over the last decades, allows a deeper understanding of the human influence on the floodplain ecosystem, and supports further studies on the impacts of forest loss and fragmentation on biodiversity, ecosystem services and human well-being in the Amazon floodplain. 1. Introduction different floodplain regions (Upper Solimões, Middle Solimões, Middle Amazon, Lower Amazon, Marajoara Gulf), each one with a specific occupation, economic and political history. This large floodplain shows an east-to-west gradient of increasing forest cover extend (Hess et al., 2015) and tree species diversity (Wittmann et al., 2006), usually assigned to natural factors such as geomorphology, rainfall seasonality and hydroperiod (DNPM, 1976; Wittmann et al., 2004, 2006). However, human activities may also play an important role in this trend, since human occupation of the floodplain progressed westward since post-Columbiam period (Cleary, 2001) contributing to habitat fragmentation, which is closely related to ecosystem integrity (Fu et al., 2013; Naeem et al., 2009; Naveh, 2007; Turner et al., 2013). Following the pace of human occupation and floristic dissimilarity, this study hypothesizes the existence of an east-to-west gradient of forest depletion caused mainly by anthropogenic factors. This hypothesis is examined through the analysis of the spatial-temporal changes of forest cover of three floodplain landscapes distributed in the upper (Landscape 1), middle (Landscape 2) and lower reaches (Landscape 3) of the Solimões/Amazonas River. This way, Landsat images acquired between the 1970’s and 2010’s were applied for mapping the state of the forest cover at the three landscapes. The maps were submitted to landscape analysis for quantifying forest loss and fragmentation over the last 40 decades. The Amazon floodplain is one of the most complex ecosystems on the planet, baring high level of biodiversity and productivity (MA, 2005; Ewel, 2010), and historically responsible for the provision of various ecosystem services to human population (MA, 2005; Ewel, 2010). Although still seen as one of the best preserved environments in the world, the Amazon floodplain has already undergone an intense process of degradation in recent decades, especially downstream Manaus due to forest loss and fragmentation (Castello et al., 2013; Renó et al., 2016, 2011). In the Lower Amazon floodplain, forest depletion is causing highly fragmented landscapes in which remnant forest patches become smaller, more susceptible to edge effects, and increasingly isolated by areas of pastures, croplands, and other anthropogenic environments (Renó et al., 2016). These changes in forest cover potentially bring a number of pervasive consequences to biodiversity and ecological processes, affecting the provision of ecosystem services to the riverine populations (Renó et al., 2016). However, forest cover changes are not continuous in space and time, since it is subject to the occupation history and to the economic and political cycles of each region. In Brazil, the main channel of Solimões/ Amazonas River runs about 3000 km from the western border to its mouth, crossing three States (Amazonas, Pará, Amapá) and five ⁎ Corresponding author. E-mail addresses: vivian.reno@inpe.br (V. Renó), evlyn.novo@inpe.br (E. Novo). https://doi.org/10.1016/j.ecolind.2018.11.019 Received 21 December 2017; Received in revised form 6 November 2018; Accepted 8 November 2018 1470-160X/ © 2018 Published by Elsevier Ltd. Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 1. Study area and landscapes: 1) São Paulo de Olivença (SPO); 2) Madeira River Mouth (MRM); and 3) Santarém (STM). and diverse. Archeological evidences show that during pre-Columbian times the floodplain was intensely occupied by larger settlements and a more numerous population than today (Denevan, 1976; Porro, 1981; Levis et al., 2017). Since the 19th century, however, floodplain land use focused activities of higher anthropogenic impact: selective logging for supplying expanding urban centers (Albernaz and Ayres, 1999; Lentini, 2005); market oriented agriculture represented by jute (Corchorus capsularis) (Winklerprins, 2006); large scale commercial fishing due to the modernization of fishing gear (eg. nylon nets, styrofoam boxes) (Almeida et al., 2003, 2001; Castello et al., 2015; Sousa and Freitas, 2011), and; deforestation (Renó et al., 2011) for the establishment of pastures and agriculture crops (Sheikh et al., 2006). At the floodplain landscape, land use pattern is adapted to the topography, which controls height and duration of inundation and, as a consequence, its biogeography (McGrath et al., 2007; Wittmann et al., 2004). Settlements and crops are concentrated on the highest terrains (levees), with communities distributed along the margins of the Amazon River and some of its tributaries (Lima, 2005; McGrath et al., 2007; Renó, 2016). Agriculture is typically carry out on the levees under private control, while fishing occurs in rivers and lakes with public access (Fraxe et al., 2007; Lima, 2005). Cattle ranching, seasonally, alternate between floodplain and upland pastures (McGrath et al., 2007; Winklerprins, 2006). Fishing is the main economic activity along the floodplain, however the Lower and Middle Amazon Regions present a longer history of economic exploitation focused on logging, agricultural and livestock activities. The economic exploitation without the proper management made these regions the most degraded landscapes of the Amazon River floodplain (Castello et al., 2013; Goulding et al., 1996; Renó et al., 2016, 2011). The study increases the scarce knowledge about the dynamics of the floodplain forest over the last decades, allows a deeper understanding of the human influence on the floodplain ecosystem, and supports further studies on the impacts of forest loss and fragmentation on biodiversity, ecosystem services and human well-being in the Amazon floodplain. 2. Study area The study area corresponds to the floodplain of Solimões-Amazon River which was subdivided into three landscapes distributed along the main channel to encompass the east-to-west gradient of anthropogenic landscape: 1) São Paulo de Olivença (SPO), located in the Upper Solimões Region; 2) Madeira River Mouth (MRM), situated near Manaus in the Middle Amazon Region, and; 3) Santarém (STM), placed in the Lower Amazon Basin 700 km from the river mouth (Fig. 1). The region has a diverse vegetation cover that includes mature forests, forests in regeneration, natural grasslands, semi-aquatic and aquatic macrophytes and crop areas (DNPM, 1978, 1977, 1976; Ferreira-Ferreira et al., 2015; McGrath et al., 2007; Winklerprins, 2006). The main ecological forcing in the floodplain is the annual flood pulse of the Amazon River (Junk et al., 1989), which shapes biota specific adaptations and controls much of the ecological and biogeochemical processes (Arraut, 2008; Junk et al., 1989; Parolin et al., 2010; Simone et al., 2003). In addition, the flood pulse influences the adaptation of human populations and the way they organize their socioeconomic activities (Lima and Pozzobon, 2005; McGrath et al., 2007). The river begins to rise between October and December (risingwater), reaching its maximum level between May and June (highwater), and falling the following months (receding-water), reaching its lowest level between September and November (low-water), depending on the region considered (west-east) (ANA, 2015). The history of human population in the Amazon floodplain is long 410 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Table 1 Landsat scenes used to create time series of forest cover between the 1970s and 2010s at SPO, MRM and STM landscapes. Date #: 1 2 3 4 5 6 7 8 SPO Date: Sensor: 1979/07 MSS-2 – – 1985/09 TM-5 1991/10 TM-5 1995/09 TM-5 1999/08 TM-5 2004/09 TM-5 2010/06 TM-5 MRM Date: Sensor: 1972 MSS-1 1981/08 MSS-2 1986/10 TM-5 1990/08 TM-5 1994/10 TM-5 1999/09 TM-5 2004/08 TM-5 2008/08 TM-5 STM Date: Sensor: 1975/08 MSS-2 1980/12 MSS-2 1987/07 TM-5 1992/09 TM-5 1997/09 TM-5 2001/11 TM-5 2004/10 TM-5 2008/09 TM-5 3. Methodology 1994). 3.1. Remote sensing data 3.4. Potential drivers and impacts Landsat optical data of MSS (Multispectral Scanner) and TM (Thematic Mapper) sensors were used to generate time series of forest cover maps between the 1970s and 2010s. For each landscape, a set of seven to eight scenes were acquired at an average interval of five years, over a total period of approximately 40 years (Table 1). The scenes were geo-referenced to a set of orthorectified images obtained from the Global Land Cover Facility database (http://glfc. umiacs.umd.edu). All maps were clipped to the floodplain area, as defined by the Amazon wetlands map produced by Hess et al. (2003) and Melack and Hess (2010), and improved by Rennó et al. (2013) and Ferreira et al. (2013). Based on the results of the landscape structure analyses and bibliographic information (Table 3), we inferred the potential drivers of forest changes at each landscape, as well as some of the likely impacts on ecosystem integrity. This step helped us to better understand the landscape dynamics of each site and to discuss the research results more consistently. To identify the potential drivers, we evaluate the dates and intensity of forest depletion in each landscape and reviewed the literature on the history of floodplain occupation, economic activities and public policies of each landscape. The authors also used official census data to assess changes economical activities in the floodplain municipalities. To infer the potential impacts of forest depletion on ecosystem integrity, the authors used the approach and the literature described in Renó et al., 2016. 3.2. Temporal mapping of forest cover Each image was classified independently, based on samples for the following classes: floodplain forest (vegetation cover dominated by tree species); water bodies (open water surface of rivers, lakes and channels); cloud (areas covered by clouds and shadows), and; unobserved (remaining types of land cover). The mapping methodology is the same used by Renó et al. (2016), and is based on a object-oriented technique that includes: a) multiresolution segmentation algorithms in a multidate approach (Boyaci et al., 2017; Desclée et al., 2006; Renó et al., 2016, 2011), and; b) Nearest Neighbor supervised classification algorithm based on fuzzy logic (Baatz and Schäpe, 2000; Boyaci et al., 2017; Renó et al., 2016, 2011). After classification, Google Earth images and pre-existing field data were used for assessing classification accuracy of the older and more recent maps of each landscape. Field data include three types of information acquired by Renó (2016) and Renó et al. (2011, 2016) in 2009 (Landscape 3) and 2014 (Landscapes 1, 2 and 3): (a) Geo-referenced photos taken by a GPS-enabled digital camera; (b) human settlement interviews, especially among the elderly, to collect information on currently and historical land cover type (around 30 years ago); and (c) botanical data (including vegetation structure, species composition and diversity) which indicated the current land cover type and helped reconstruct its approximate evolution over the past ∼30 years. Based on these data, confusion matrices were built and used to estimate classification accuracy through calculation of the Kappa index of agreement. The overall accuracy of the maps ranged from 0.80 to 0.75 among the most ancient and recent maps respectively. 4. Results 4.1. Habitat Results show pronounced differences in forest loss among the three landscapes, especially between SPO and STM. While SPO shows a reduction of only 1.3% (4510 ha) of the forest habitat between 1970s and 2010s, MRM and STM show 29% (92,626 ha) and 70% (196,535 ha) of forest loss, respectively (Fig. 2a). Between 1970s and 2010s, the area of the largest forest fragment decreased only 2.7% in SPO, but 26% in MRM and 89% in STM (Fig. 2a – largest patch). Once again STM presents the greatest change. The largest patch represented 58.5% of the total forest area in 1970s at STM landscape, decreasing to only 1.9% of the total forest area in 2010s. In addition, only 62% of STM remaining forest (2010s) represents primary forest, in contrast to 81% and 97% of primary forest remnants in MRM and SPO, respectively. Time-series of forest cover maps show that forest dynamic also differs among the landscapes (Fig. 3a). SPO shows great forest cover stability over the analyzed period, with the greatest alterations in 1991–2004 (dates # 4–7). MRM presents a gradual forest loss over the period, being higher in 1972–1981 (dates # 1–2), with a slight increase in 1981–1986 (dates # 2–3). In contrast, STM shows an older and more pronounced forest dynamic, with larger losses in 1980–1987 (dates # 2–3) and 1992–1997 (dates # 4–5). 3.3. Landscape analysis 4.2. Anthropization The landscape analysis was based on a set of indicators, divided in four categories of forest depletion: Habitat, Anthropization, Fragmentation and Connectivity (Table 2). The definition of those indicators was based on literature addressing landscape fragmentation from the perspective of ecological groups relevant to the floodplain region (Renó et al., 2016; Renó, 2016). Metric computations were performed in ArcGis (ESRI, 2012) and Fragstats (McGarigal and Marks, Deforestation data show the same pattern as the other indicators, being lower at SPO (4.1% of the total area), intermediate at MRM (30.1% of the total area), and higher at STM landscape (51% of the total area) (Fig. 4a). Data on secondary deforestation, however, show large similarity among landscapes. For all landscapes, secondary deforestation corresponds to less than 4% of the total deforested area (Fig. 4b). The periods of higher deforestation rates in SPO are 1995–1999 411 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Table 2 Forest depletion indicators for the analysis of landscape dynamic. Indicators Description Unit Threshold Habitat Forest area1 Largest patch Primary forest2 Area and percent of the landscape comprised by forest Area and percent of the forest cover comprised by the largest patch Percent of forest area comprised by primary forest. ha/% ha/% % Edge width (m): 100 Edge width (m): 100 Edge width (m): 100 Anthropization Deforestation area3 Secondary deforestation4 Area and percent of the landscape comprised by deforestation. Percent of deforested area comprised by secondary deforestation. ha/% % Edge width (m): 100 Edge width (m): 100 Fragmentation Mean patch size/number of patches Edge area Mean size and total number of forest patches Area and percentage of forest submitted to edge effects for different edge widths ha/− ha/% Edge width (m): 0 Edge width (m): 100 Connectivity Connectivity index5 Mean proximity index6 Proportion of functional joins between all forest patches considering different distances Mean Euclidean distance among patches that are within a specific search radius % – Distance (m): 100, 500, 1000 Search radius (m): 100, 500, 1000 1 2 3 4 5 6 Primary and regrowth forest. Intact forest throughout the time series. Deforestation accumulated throughout the time series. Deforested more than once throughout the time series. Equals 0 when the landscape is composed of a single patch or none of the forest patches are connected; equals 100 when all patches are connected. Increases as patches become closer and are more contiguous or less fragmented in distribution and vice versa (unit less). (dates # 1–4), with a great peak in 1980–1987 (dates # 2–3), and gradual decrease over the remainder period (Fig. 3b). Deforestation at SPO is concentrated in islands and borders of the Solimões River (Fig. 5). In contrast, at MRM and STM the deforestation spread in the entire landscapes. In MRM deforested areas are concentrated in the transition between floodplain and upland, and also at water bodies' margins (Fig. 6). However, at STM, deforested areas are located especially on water bodies' margins (lakes, rivers and canals), higher terrains, in the transition between floodplain and upland, and near urban centers (Fig. 7). Table 3 Bibliographic references used to infer the potential drivers of forest changes and the impacts on ecosystem services at each landscape. Drivers of forest changes Alencar 2005; Blackman et al., 2017; Espinoza et al., 2013; IBGE, 2015; Lima and Pozzobon, 2005; Marengo and Espinoza, 2016; Nepstad et al., 2006; Pantoja, 2005; Silva et al., 2001; Winklerprins, 2006. Impacts on ecosystem integrity Balmford and Bond, 2005; Brown, 1997; Brown and Albrecht, 2001; Didham et al., 1996; Dirzo and Raven, 2003; Dohm et al., 2011; Fahrig, 2003; Fu et al., 2013; Guimarães et al., 2014; Hammond and Miller, 1998; Kerr et al., 2001; Kremen, 1992; Laurance et al., 2002; Lees and Peres, 2009; Lovejoy et al., 1986; Naeem et al., 2009; Naveh, 2007; Osborn et al., 1999; Powell and Powell, 1987; Renó et al., 2016; Ricketts et al., 2008, 2006; Santos-Filho et al., 2012; Stratford and Stouffer, 1999; Turner et al., 2013; Urbas et al., 2007. 4.3. Fragmentation Forest fragmentation is also the smallest at SPO and the largest at STM landscape. In this case, however, the differences between SPO and the other landscapes are not as great as observed previously (Fig. 8). In SPO, the number of forest fragments increased 31% along the period, while in MRM and STM forest fragments increased 45% and 79% respectively. The increase in the number of forest fragments was followed by 25% decrease of their mean size in SPO, 51% in MRM and 83% in STM. Forest fragmentation creates new forest edges. Therefore, in the last 40 years there was a considerable increase in edge areas for all (dates # 5–6) and 1999–2004 (dates # 6–7), while the periods of lower deforestation rates are 1985–1991 (dates # 3–4) and 2004–2010 (dates # 7–8). In MRM, deforestation rate is higher in is 1972–1981 (dates # 1–2), falling considerably in 1981–1986 (dates # 2–3), rising again in 1986–1999 (dates # 3–6), and falling once again in 1999–2008 (dates # 6–8). In STM, the period of highest deforestation rate is 1975–1992 Fig. 2. Indicators of habitat in SPO, MRM and STM landscapes: a) Forest area and largest patch decrease between 1970s and 2010s; b) Primary forest area throughout the period. 412 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 3. Forest cover (a) and deforestation (b) evolution between 1970s and 2010s in SPO, MRM and STM landscapes. 1Dates 1 to 8 of each landscape are specified in Table 1. 2Deforestation on date 1 cannot be estimated (base date). stability over the period; especially for functional distances of 100 m. Mean proximity index also indicate high forest cover stability in SPO, with an average decrease of only 4% over the period, ranging from 3.5% to 5.2% depending on the functional distance (100 m, 500 m, 1 km) (Fig. 9b). MRM shows a higher decrease in connectivity index, averaging 44% and ranging from 36% to 54%. The values were extremely low, not exceeding 0.3% in 1970s and 0.2% in 2010s (Fig. 9a). Compared to SPO, patch connectivity in the MRM is smaller, indicating larger forest degradation. However, given that the maximum connectivity in the region is unknown, the time changes in connectivity values are more important than the metric value itself. Mean proximity also presented considerable changes in MRM, but with low difference among search radii. Between 1970s and 2000s, there was an average decrease of 55%, ranging from 54.9% to 55.5% (Fig. 9b). STM shows great alterations in connectivity index, with an average decreased of 63% along the period (Fig. 9a). Considering different functional distances (100 m, 500 m and 1 km), the reduction varied from 76% to 50%. The value of the index indicates that patch connectivity was already low in 1970s, not exceeding 0.4%. However, as mentioned previously, it is not possible to attribute this low value only to forest cover degradation. Mean proximity index presented even greater alterations in STM, with a decrease of more than 99% for all the landscapes, except for SPO (Fig. 8). Forest edges increased only 3.9% in SPO, but 30% in MRM and 74% in STM landscape. Due to the large reduction of forest cover in most of the landscapes, the increase of edge areas is reported only as percentage of forest. In 1970 s, the edges represented 38% of the forest area in STM, coming to represent 66% in 2010s. In MRM, the edges represented 39% of the forest area in 1970s and 50% in 2010s. In SPO, however, the edges represented only 17% of the forest area in 1970s and 18% in 2010s. 4.4. Connectivity Connectivity also varied among the landscapes, following a similar pattern than observed in previous indicators (SPO – better condition, MRM – intermediate condition, STM – worst condition). Connectivity index in SPO decreased by an average of 22% between 1970s and 2010s, ranging from 16% to 28% depending on the functional distance (100 m, 500 m, 1 km) (Fig. 9a). It is important to highlight, however, that the connectivity of a floodplain landscape will never reach 100% due to the large proportion of water bodies. Although the maximum connectivity expected for this landscape is unknown, it can be assumed that it is close to the observed values, since SPO forest cover showed little change in the last decades, being dominated by primary forest (72%). In this case, results show high patch connectivity, with great Fig. 4. Indicator of anthropization in SPO, MRM and STM landscapes: a) Deforestation; b) Secondary deforestation. 413 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 5. Map of forest cover change in SPO landscape between 1970s and 2010s. Environment and Renewable Natural Resources (IBAMA). As a consequence, in the 1980s, fishing became the main economic activity of the riverine population, with high productivity during the catch period of some species, such as catfish. These facts may be related to the stability and slight increase in forest cover between 1979 and 1991 (Fig. 10a). In fact, some studies suggest that the presence of indigenous reservation acts as a mechanism to contain deforestation, not only inside but also outside the reservation perimeter (Blackman et al., 2017; Nepstad et al., 2006). Indigenous lands occupy about 10% of this landscape and encompass 9.5% of the 2010 forest area. It may seem like a small portion, but much of the reserve area is locate on the adjacent uplands and on the floodplains east of the study area. In all, there are nine indigenous lands in the region, which together total 532,805 ha of area, of which 45,617 ha are in the area of SPO landscape. In the early 1990s there was a large migration of the floodplain population to urban and rural areas due to economic changes and environmental and social factors, such as the scarcity of fishery resources related to overfishing, extreme floods, and the lack of public services and assistance. In the 2000s, however, the lack of job opportunities forced some families to return to the floodplain (Alencar, 2005). The waves of migration and return of the floodplain population seem to be related to the alternation between decrease and subsequent increase of forest cover in 1991 and 2004, respectively. In this case, it is hypothesized that the presence of traditional floodplain population is beneficial to forest integrity, probably due to the inhibition of predatory forest exploitation by outsiders (Lima and Pozzobon, 2005; Silva et al., 2001). An alternative hypothesis, which does not necessarily exclude the previous one, is the destruction of part of the forest cover due to the extreme floods of 1989 and 1999 (Espinoza et al., 2013; Marengo and Espinoza, 2016), with subsequent forest regeneration from 2004 on search radii (Fig. 9b). 5. Discussion Results show a gradual increase in the fragmentation process of floodplain forest from the western to the eastern landscapes of the Solimões/Amazonas River. The data also indicate larger similarity between STM and MRM landscapes, and a great difference between these two and SPO landscape. SPO presented almost no changes in landscape structure compared to the other landscapes. Forest loss was minimal, as was the reduction of riparian forest, the decrease of the largest patch area, the increase of edge areas and changes in the mean proximity of patches. Besides temporal stability, SPO landscape is dominated by floodplain forest, mostly primary. The most significant alterations in SPO landscape are related to the increase in number of patches, the decrease in their mean size and the loss of connectivity. Different authors (Fu et al., 2013; Naeem et al., 2009; Naveh, 2007; Turner et al., 2013) report that the structure of a given landscape is closely related to the integrity of its ecosystems and, therefore, to its capacity to provide ecosystem services. In this context, results of SPO landscape indicate high integrity of the forest ecosystem and high capacity for provision of ecosystem services that are dependent on the integrity of the forest habitat. The periods of stability and alteration of forest cover in SPO may be related to the socioeconomic processes that occurred in the region. According to Alencar (2005), the period between 1970 and 1980 is marked by the decline of timber production in this landscape, due to enforcement of new public policies including the demarcation of indigenous reservation and forest conservation legislation under the control of a new federal agency: the Brazilian Institute of the 414 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 6. Map of forest cover change in MRM landscape between 1970 s and 2010 s. Bond, 2005; Dirzo and Raven, 2003). In the Amazon, forest conversion can lead to irreversible shifts in the composition and diversity of important insects, such as bees (Kerr et al., 2001) and Lepidoptera groups (Brown, 1997). Some of these organisms are so important that are often used as indicators of the environmental equilibrium (Brown, 1997; Kremen, 1992) and of the biodiversity of terrestrial invertebrates (Hammond and Miller, 1998; Kremen, 1992; Osborn et al., 1999). Results of STM also show a high degree of forest fragmentation and great loss of patch connectivity over the period. In fact STM presented great reduction of the mean patch size, a considerable increase in edge area, and great decrease in the mean proximity of patches. The reduction of the patch area increases tree species mortality and reduces the richness and diversity of animal and plant species, which, in turn, negatively impact the provision of many ecosystem services (Fahrig, 2003; Laurance et al., 2002). The increase of edge areas severely affect the microclimatic conditions, impacting several ecosystem process, such as the mortality and diversity of trees, birds and mammals (Laurance et al., 2002; Lovejoy et al., 1986; Santos-Filho et al., 2012), and the balance of herbivores and parasitoids communities that are important pollinators and act as natural biological control (Didham et al., 1996; Dohm et al., 2011; Guimarães et al., 2014; Laurance et al., 2002; Lovejoy et al., 1986; Urbas et al., 2007). The loss of forest connectivity is also related to several pervasive effects. Some of them are the decrease of tree diversity and richness (Metzger, 2000), the decrease of bird diversity and bird movement on the landscape (Lees and Peres, 2009; Stratford and Stouffer, 1999), the decrease of mammal richness (Lees and Peres, 2008), the decrease of pollinator insect richness, diversity and visitation rates (Brown and Albrecht, 2001; Powell and (Fig. 10a). In contrast, STM was the landscape with the largest forest cover change in the period. Results show disturbing changes in the landscape integrity, including drastic reduction of forest cover. The periods of extreme forest cover reduction (1980–1987 and 1992–1997; dates # 2–3 and 4–5) coincide with a short period of jute market recovery and the introduction of cattle raising, initially buffalo in the STM region. Jute production was a major driver of floodplain deforestation in the beginning of the twentieth century, but the loss of jute market in the 1960s gave space for crop abandonment, followed for forest regrowth (Winklerprins, 2006). The author, however, points out that despite the general trend of diminishing jute production, as of 1965 there were two peaks of significant increase in production between 1980 and 1987, specifically in 1982 and 1986. Conversely, the period between 1975 and 1980 was marked by an intense decline in jute production (Winklerprins, 2006), what supports the relative stability of forest cover reported in this study. As of 1986, there was a large decrease in jute production and buffalo ranching came to play a more important economic role in the region (Winklerprins, 2006). In fact, data on buffalo herds (IBGE, 2015) show that the greatest increase in the region occurred between 1990 and 1995, coinciding with the second period of intense forest loss (1992–1997; dates # 4–5) reported in the present study (Fig. 10c). Unlike SPO landscape, changes detected in the landscape structure of STM points to serious ecological implications, suggesting huge threats to biodiversity and, as a consequence, to the provision of ecosystem services (Renó et al., 2016). Habitat loss is the main cause of biodiversity decline at local, regional and global scale (Balmford and 415 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 7. Map of forest cover change in STM landscape between 1970s and 2010s. impacts on the floodplain ecosystem functioning, since the similarities between MRM and STM landscapes indicate similar degradation processes, albeit to a less severe degree and/or a less advanced stage. The periods of greater forest cover variation in MRM seem to be related to agricultural activities and livestock expansion in the region. According to Pantoja (2005), the Middle Amazon region was part of the jute cycle, whose production declined in 1965. However, data from Winklerprins (2006) show that the decline was not linear, with great oscillations in jute production between 1965 and 1986. The greatest one occurred between 1970 and 1975, when there was a high increase in jute production (∼58,000 tons) that almost reached the amount produced at the peak of the cycle (∼62,000 tons). This fact may be related to the high rate of deforestation between 1972 and 1980. Another hypothesis is the expansion of cattle ranching in the floodplain of MRM region. According to Pantoja (2005) the expansion started in 1970. However, data on bovine/buffalo herds for the municipalities that compose the landscape show a significant increase only from 1986 (IBGE, 2015), explaining the gradual decrease in forest cover only as of this date. According to these facts, the region underwent an economic transition between 1980 and 1986, with relative reduction of anthropogenic impacts on forest cover. This may have allowed the temporary regeneration of part of the forest habitat, explaining the forest cover trend reported in this study for the period between 1981 and 1986 (dates # 2–3) (Fig. 10b). The comparison of landscapes shows that the eastern landscape (STM) presented an older e more intense degree of forest depletion, unlike the western landscape (SPO) that presented a much more recent Fig. 8. Indicators of fragmentation in SPO, MRM and STM landscapes: % decrease in mean patch size, % increase in number of patches and, % increase in edge areas between 1970s and 2010s. Powell, 1987; Ricketts et al., 2008, 2006), and others (Kruess and Tscharntke, 2000, 1994). In relation to MRM landscape, it presented intermediate condition of forest cover degradation compared to the other landscapes, with characteristics more similar to STM. Alterations of forest cover, edge areas and mean proximity are the most concern changes. The data suggest that MRM landscape is also subject to strong anthropogenic 416 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo Fig. 9. Indicators of connectivity in SPO, MRM and STM landscapes: a) Connectivity and; b) Mean proximity decrease between 1970s and 2010s. Fig. 10. Relationship between forest cover changes and possible drivers in: a) SPO; b) MRM and; c) STM landscapes. 417 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo and less intense degradation process. Compared to the others, the middle landscape (MRM) presented an intermediate condition of time and intensity of forest depletion, although being more similar to eastern (STM) than to western (SPO) landscape. The main drivers of forest cover variation reported for the western landscape (SPO) seems to be the demarcation of indigenous lands and the creation of new policies on timber that lead to shift on the main economic activity, from timber to fishing. In the other landscapes however, the main factors responsible for forest cover changes were jute production and cattle/buffalo ranching. Although the main economic activity of the Amazon floodplain is fishing, the middle and lower regions have a longer history of economic exploitation focused on agricultural and livestock activities, probably due to its ease access compared to the upper region. Informationsverarbeitung. Proceedings, Salzburg, pp. 12–23. Balmford, A., Bond, W., 2005. Trends in the state of nature and their implications for human well-being. Ecol. Lett. 8, 1218–1234. https://doi.org/10.1111/j.1461-0248. 2005.00814.x. Blackman, A., Corral, L., Lima, E.S., Asner, G.P., 2017. Titling indigenous communities protects forests in the Peruvian Amazon. Proc. Natl. Acad. Sci. 114, 4123–4128. https://doi.org/10.1073/pnas.1603290114. Boyaci, D., Erdogan, M., Yildiz, F., 2017. Pixel- versus object-based classification of forest and agricultural areas from multiresolution satellite images. Turk. J. Electr. Eng. Comput. Sci. 25, 365–375. https://doi.org/10.3906/elk-1504-261. Brown, C., Albrecht, C., 2001. The effect of tropical deforestation on stingless bees of the Genus Melipona (Insecta: Hymenoptera: Apidae: Meliponini) in Central Rondonia, Brazil. J. Biogeogr. 28, 623–634. https://doi.org/10.1046/j.1365-2699.2001. 00583.x. Brown, K.S., 1997. Diversity, disturbance, and sustainable use of Neotropical forests: insects as indicators for conservation monitoring. J. Insect Conserv. 1, 25–42. https:// doi.org/10.1023/A:1018422807610. Castello, L., Isaac, V.J., Thapa, R., 2015. Flood pulse effects on multispecies fishery yields in the Lower Amazon. R. Soc. Open Sci. 2, 150299. https://doi.org/10.1098/rsos. 150299. Castello, L., McGrath, D.G., Hess, L.L., Coe, M.T., Lefebvre, P.A., Petry, P., Macedo, M.N., Renó, V.F., Arantes, C.C., 2013. The vulnerability of Amazon freshwater ecosystems. Conserv. Lett. 6, 217–229. https://doi.org/10.1111/conl.12008. Cleary, D., 2001. Towards an environmental history of the Amazon: from prehistory to the Nineteenth Century. Lat. Am. Res. Rev. 36, 65–96. Denevan, W., 1976. The Native Population of the Americas in 1492. University of Winsconsin Press, Madison. Desclée, B., Bogaert, P., Defourny, P., 2006. Forest change detection by statistical objectbased method. Remote Sens. Environ. 102, 1–11. Didham, R.K., Ghazoul, J., Stork, N.E., Davis, A.J., 1996. Insects in fragmented forests: a functional approach. Trends Ecol. Evol. 11, 255–260. https://doi.org/10.1016/01695347(96)20047-3. Dirzo, R., Raven, P.H., 2003. Global state of biodiversity and loss. Annu. Rev. Environ. Resour. 28, 137–167. https://doi.org/10.1146/annurev.energy.28.050302.105532. DNPM, 1978. PROJETO RADAMBRASIL – Folha SA. 20 Manaus – Geologia, Geomorfologia, Pedologia, Vegetação, Uso Potencial da Terra, (Levantamentos de Recursos Naturais). DNPM, Rio de Janeiro. DNPM, 1977. PROJETO RADAMBRASIL – Folha SA. 19 Içá – Geologia, Geomorfologia, Pedologia, Vegetação, Uso Potencial da Terra, (Levantamentos de Recursos Naturais). DNPM, Rio de Janeiro. DNPM, 1976. PROJETO RADAMBRASIL – Folha SA. 21 Santarém – Geologia, Geomorfologia, Pedologia, Vegetação, Uso Potencial da Terra, (Levantamentos de Recursos Naturais). DNPM, Rio de Janeiro. Dohm, C., Leal, I.R., Tabarelli, M., Meyer, S.T., Wirth, R., Dohm, C., Leal, I.R., Tabarelli, M., Meyer, S.T., Wirth, R., 2011. Leaf-cutting ants proliferate in the Amazon: an expected response to forest edge? J. Trop. Ecol. 27, 645–649. https://doi.org/10. 1017/S0266467411000447. Espinoza, J.C., Ronchail, J., Frappart, F., Lavado, W., Santini, W., Guyot, J.L., 2013. The major floods in the Amazonas River and Tributaries (Western Amazon Basin) during the 1970–2012 period: a focus on the 2012 flood. J. Hydrometeorol. 14, 1000–1008. https://doi.org/10.1175/JHM-D-12-0100.1. ESRI, 2012. ArcGIS Desktop. Ewel, K.C., 2010. Appreciating tropical coastal wetlands from a landscape perspective. Front. Ecol. Environ. 8, 20–26. https://doi.org/10.1890/080090. Fahrig, L., 2003. Effects of habitat fragmentation on biodiversity. Rev. Lit. Arts Am. 34, 487–515. https://doi.org/10.1146/132419. Ferreira-Ferreira, J., Silva, T.S.F., Streher, A.S., Affonso, A.G., de Almeida Furtado, L.F., Forsberg, B.R., Valsecchi, J., Queiroz, H.L., de Moraes Novo, E.M.L., 2015. Combining ALOS/PALSAR derived vegetation structure and inundation patterns to characterize major vegetation types in the Mamirauá Sustainable Development Reserve, Central Amazon floodplain, Brazil. Wetlands Ecol. Manage. 23, 41–59. Ferreira, R.D., Leão, J.A.D., Silva, T.S.F., Rennó, C.D., Novo, E.M.L.M., Barbosa, C.C.F., 2013. Atualização e correção do delineamento de áreas alagáveis para a bacia Amazônica. In: XVI Simpósio Brasileiro de Sensoriamento Remoto. Instituto Nacional de Pesquisas Espaciais (INPE), Foz do Iguaçu, PR, pp. 5864–5871. Fraxe, T., Pereira, H., Witkoski, A., 2007. Comunidades ribeirinhas amazônicas: modos de vida e uso dos recursos naturais. Fu, B., Wang, S., Su, C., Forsius, M., 2013. Linking ecosystem processes and ecosystem services. Curr. Opin. Environ. Sustain. 5, 4–10. https://doi.org/10.1016/j.cosust. 2012.12.002. Goulding, M., Smith, N., Mahar, D., 1996. Floods of Fortune: Ecology and Economy along the Amazon. Columbia University Press, New York. Guimarães, C., Viana, J.P.R., Cornelissen, T., 2014. A meta-analysis of the effects of fragmentation on herbivorous insects. Environ. Entomol. 43, 537–545. https://doi. org/10.1603/EN13190. Hammond, P.C., Miller, J.C., 1998. Comparison of the biodiversity of lepidoptera within three forested ecosystems. Ann. Entomol. Soc. Am. 91, 323–328. https://doi.org/10. 1093/aesa/91.3.323. Hess, L.L., Melack, J.M., Novo, E.M.L.M., Barbosa, C.C.F., 2003. Dual-season mapping of wetland inundation and vegetation. Remote Sens. Environ. 87, 404–428. Hess, L.L., Melack, J.M., Affonso, A.G., Barbosa, C.C.F., Gastil-Buhl, M., Novo, E.M.L.M., 2015. Wetlands of the Lowland Amazon Basin: extent, vegetative cover, and dualseason inundated area as mapped with JERS-1 synthetic aperture radar. Wetlands 35 (4), 745–756. IBGE – Brazilian Institute of Geography and Statistics (IBGE), 2015. Agricultural Census 1940-2006. Instituto Brasileiro de Geografia e Estatística (IBGE). 6. Conclusions The present study shows the existence of an east-to-west gradient of time and intensity of forest depletion along the floodplain of the Solimões/Amazonas River, and provides evidence that the gradient is a response to human occupation process and public policies either protecting or stimulating changes in land use. Forest depletion gradient follows the pace of human occupation, and also the pattern of forest cover extent and tree diversity along the floodplain, indicating that these floristic differences also have a strong anthropogenic component. The study warns to the accelerated process of forest degradation on the eastern regions of the Amazon floodplain, which has great potential for depleting the biodiversity and ecological processes (Renó et al., 2016). In this context, the next steps for this research include combining landscape data with: a) ecological in situ data such floristic and animal inventories; b) interview data on the perception of ecosystem service provision; and, c) interview data on the well-being of riverine populations. These approaches are already being applied and will permit the assessment of the impacts of forest depletion on biodiversity, ecosystem services and human well-being in the Amazon floodplain. Acknowledgments The authors wish to acknowledge the São Paulo Research Foundation (FAPESP/Brazil) and the Brazilian National Council for Scientific and Technological Development (CNPq) for the financial support during the field campaigns (CNPq 301276/2010-2; FAPESP 2011/23594-8) Research Productivity fellowship support Evlyn Novo (CNPq 304568/2014-7) and for the fellowships supporting Vivian Renó through her Ph.D. (Process FAPESP Process 2012/02544-5) and postdoctoral research program (Process CNPq 300604/2017-3). The authors also thank the anonymous reviewers for their valuable comments on the manuscript. References Albernaz, A.K.M., Ayres, J.M., 1999. Selective logging along the middle Solimoes river. In: Padoch, C. (Ed.), Várzea: Diversity, Development, and Conservation of Amazonia’s Whitewater Floodplains. Botanical Garden Press, New York, USA, pp. 135–151. Alencar, E.F., 2005. Políticas públicas e (in)sustentabilidade social: o caso de comunidades de várzea no Alto Solimões, Amazonas. In: Lima, D. (Ed.), Diversidade Socioambiental Nas Várzeas Dos Rios Amazonas E Solimões: Perspectivas Para O Desenvolvimento Da Sustentabilidade. Ibama, ProVárzea, Manaus, AM, pp. 59–100. Almeida, O.T., Lorenzen, K., Mcgrath, D.G., 2003. Commercial fishing in the Brazilian Amazon: regional differentiation in fleet characteristics and efficiency. Fish. Manage. Ecol. 10, 109–115. https://doi.org/10.1046/j.1365-2400.2003.00320.x. Almeida, O.T., McGrath, D.G., Ruffino, M.L., 2001. The commercial fisheries of the lower Amazon: an economic analysis. Fish. Manage. Ecol. 8, 253–269. https://doi.org/10. 1046/j.1365-2400.2001.00234.x. ANA – National Water Agency Brazil (ANA), 2015. HidroWeb [WWW Document]. Fluviom. time Ser (accessed 7.1.15). Arraut, E.M., 2008. Migração do peixe-boi amazônico: uma abordagem por sensoriamento remoto, radiotelemetria e geoprocessamento. Instituto Nacional de Pesquisas Espaciais. Baatz, M., Schäpe, A., 2000. Multiresolution Segmentation - an optimization approach for high quality multi-scale image segmentation. In: XII Angewandte Geographische 418 Ecological Indicators 98 (2019) 409–419 V. Renó, E. Novo vida. In: Lima, D. (Ed.), Diversidade Socioambiental Nas Várzeas Dos Rios Amazonas E Solimões: Perspectivas Para O Desenvolvimento Da Sustentabilidade. Ibama, ProVárzea, Manaus, AM, pp. 157–206. Parolin, P., Wittmann, F., Schöngart, J., 2010. Tree phenology in Amazonian floodplain forests. Central Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management. Springer, Berlin/Heidelberg/New York. Porro, A., 1981. O Povo das águas: ensaios de etno-história amazônica. Vozes/Edusp, Petrópolis. Powell, A., Powell, G., 1987. Population dynamics of male euglossine bees in Amazonian forest fragments. Biotropica. https://doi.org/10.2307/2388742. Rennó, C.D., Novo, E.M.L.M., Banon, L.C., 2013. Correção geométrica da máscara de áreas alagáveis da bacia amazônica. In: XVI Simpósio Brasileiro de Sensoriamento Remoto. Instituto Nacional de Pesquisas Espaciais (INPE), Foz do Iguaçu, PR, pp. 5507–5514. Renó, V., Novo, E., Escada, M., 2016. Forest fragmentation in the Lower Amazon Floodplain: implications for biodiversity and ecosystem service provision to riverine populations. Remote Sens. 8, 886. https://doi.org/10.3390/rs8110886. Renó, V.F., 2016. As várzeas amazônicas: alterações da paisagem e seus impactos na integridade florística, provisão de serviços ecossistêmicos e bem-estar de comunidades ribeirinhas. Instituto Nacional de Pesquisas Espaciais (INPE). Renó, V.F., Novo, E.M.L.M., Suemitsu, C., Rennó, C.D., Silva, T.S.F., 2011. Assessment of deforestation in the Lower Amazon floodplain using historical Landsat MSS/TM imagery. Remote Sens. Environ. 115, 3446–3456. https://doi.org/10.1016/j.rse. 2011.08.008. Ricketts, T.H., Regetz, J., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Bogdanski, A., Gemmill-Herren, B., Greenleaf, S.S., Klein, A.M., Mayfield, M.M., Morandin, L.A., Ochieng’, A., Viana, B.F., 2008. Landscape effects on crop pollination services: are there general patterns? Ecol. Lett. 11, 499–515. https://doi.org/10.1111/j.14610248.2008.01157.x. Ricketts, T.H., Williams, N.M., Mayfield, M.M., 2006. Connectivity and ecosystem services: crop pollination in agricultural landscapes. In: Crooks, K.R., Sanjayan, M.A. (Eds.), Connectivity Conservation. Cambridge University Press, Cambridge, U.K., pp. 255–290. Santos-Filho, M., Peres, C.A., Silva, D.J., Sanaiotti, T.M., 2012. Habitat patch and matrix effects on small-mammal persistence in Amazonian forest fragments. Biodivers. Conserv. 21, 1127–1147. https://doi.org/10.1007/s10531-012-0248-8. Sheikh, P.A., Merry, F.D., McGrath, D.G., 2006. Water buffalo and cattle ranching in the Lower Amazon Basin: comparisons and conflicts. Agric. Syst. 87, 313–330. https:// doi.org/10.1016/j.agsy.2005.02.003. de Silva, D. de F.L., Santos, A.K.C.R. dos, Santos, S.E.B. dos, 2001. Diversidade genética de populações humanas na Amazônia. In: Vieira, I., Silva, J., Oren, D., D’Incao, M. (Eds.), Diversidade Cultural E Biológica Da Amazônia. Museu Emílio Goeldi, Belém, pp. 165–193. Simone, O. De, Haase, K., Muller, E., Junk, W.J., Hartmann, K., Schreiber, L., Schmidt, W., 2003. Apoplasmic barriers and oxygen transport properties of hypodermal cell walls in roots from four Amazonian tree species. Plant Physiol. 132, 206–217. Sousa, R.G.C., Freitas, C.E. de C., 2011. Seasonal catch distribution of tambaqui (Colossoma macropomum), Characidae in a central Amazon floodplain lake: implications for sustainable fisheries management. J. Appl. Ichthyol. 27, 118–121. https:// doi.org/10.1111/j.1439-0426.2010.01521.x. Stratford, J.A., Stouffer, P.C., 1999. Local extinctions of terrestrial insectivorous birds in a fragmented landscape near Manaus, Brazil. Conserv. Biol. 13, 1416–1423. https:// doi.org/10.1046/j.1523-1739.1999.98494.x. Turner, M.G., Donato, D.C., Romme, W.H., 2013. Consequences of spatial heterogeneity for ecosystem services in changing forest landscapes: priorities for future research. Landsc. Ecol. 28, 1081–1097. https://doi.org/10.1007/s10980-012-9741-4. Urbas, P., Araújo Jr., M.V., Leal, I.R., Wirth, R., 2007. Cutting more from cut forests: edge effects on foraging and herbivory of leaf-cutting ants in Brazil. Biotropica 39, 489–495. https://doi.org/10.1111/j.1744-7429.2007.00285.x. Winklerprins, A.M.G.A., 2006. Jute cultivation in the lower Amazon, 1940–1990: an ethnographic account from Santarém, Pará, Brazil. J. Hist. Geogr. 32, 818–838. https://doi.org/10.1016/j.jhg.2005.09.028. Wittmann, F., Junk, W.J., Piedade, M.T.F., 2004. The várzea forests in Amazonia: flooding and the highly dynamic geomorphology interact with natural forest succession. For. Ecol. Manage. 196, 199–212. https://doi.org/10.1016/j.foreco.2004.02. 060. Wittmann, F., Schöngart, J., Montero, J.C., Motzer, T., Junk, W.J., Piedade, M.T.F., Queiroz, H.L., Worbes, M., 2006. Tree species composition and diversity gradients in white-water forests across the Amazon Basin. J. Biogeogr. 33, 1334–1347. Junk, W.J., Bayley, P.B., Sparks, R.E., 1989. The Flood Pulse Concept in River-Floodplain Systems. Int. Large River Symp. Kerr, W.E., Carvalho, G.A., Silva, A.C., Assis, M.G.P., 2001. Aspectos pouco mencionados da biodiversidade amazônica. Mensagem Doce 80, 45–60. https://doi.org/10.1590/ S0102-311X1989000100011. Kremen, C., 1992. Assessing the indicator porperties of species assemblages for natural areas monitoring. Ecol. Appl. 2, 203–217. https://doi.org/10.2307/1941776. Kruess, A., Tscharntke, T., 2000. Species richness and parasitism in a fragmented landscape: experiments and field studies with insects on Vicia sepium. Oecologia 122, 129–137. https://doi.org/10.1007/PL00008829. Kruess, A., Tscharntke, T., 1994. Habitat fragmentation, species loss, and biological control. Science 80 (264), 1581–1584. https://doi.org/10.1126/science.264.5165. 1581. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L., Bruna, E.M., Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G., Sampaio, E., 2002. Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conserv. Biol. 16, 605–618. https://doi.org/10.1046/j.1523-1739.2002.01025.x. Lees, A.C., Peres, C.A., 2008. Conservation value of remnant riparian forest corridors of varying quality for Amazonian birds and mammals. Conserv. Biol. 22, 439–449. https://doi.org/10.1111/j.1523-1739.2007.00870.x. Lees, A.C., Peres, C.A., 2009. Gap-crossing movements predict species occupancy in Amazonian forest fragments. Oikos 118, 280–290. https://doi.org/10.1111/j.16000706.2008.16842.x. Lentini, M., 2005. Fatos Florestais da Amazônia. Instituto do Homem e Meio Ambiente da Amazônia. Levis, C., Costa, F.R.C., Bongers, F., Peña-claros, M., Clement, R., Junqueira, A.B., Neves, E.G., Tamanaha, E.K., Figueiredo, F.O.G., Salomão, R.P., Castilho, C.V., William, E., Phillips, O.L., Guevara, J.E., Sabatier, D., 2017. Persistent effects of pre-Columbian plant domestication on Amazonian forest composition. Science (80-) 355, 925–931. https://doi.org/10.1126/science.aal0157. Lima, D., 2005. Diversidade socioambiental nas várzeas dos rios Amazonas e Solimões: perspectivas para o desenvolvimento da sustentabilidade. Ibama, ProVárzea, Manaus, AM. Lima, D., Pozzobon, J., 2005. Amazônia socioambiental: sustentabilidade ecológica e diversidade social. Estud. avançados 19, 45–76. Lovejoy, T.E., Bierregaard, R.O., Rylands, A.B., Malcom, J.R., Quintela, C.E., Harper, L.H., Brown, K.S., Powell, A.H., Powell, G.V.N., Shubart, H.O.R., Hays, M.B., 1986. Edge and other effects of isolation on Amazon forest fragments. Conserv. Biol. Sci. Scarcity Divers. 257–325. MA – Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being: Synthesis – Wetlands and Water. Island Press, Washington DC. Marengo, J.A., Espinoza, J.C., 2016. Extreme seasonal droughts and floods in Amazonia: causes, trends and impacts. Int. J. Climatol. 36, 1033–1050. https://doi.org/10. 1002/joc.4420. McGarigal, K., Marks, B., 1994. FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. Corvallis, OR. McGrath, D.G., Almeida, O.T., Merry, F.D., 2007. The influence of community management agreements on household economic strategies: Cattle grazing and fishing agreements on the Lower Amazon floodplain. Int. J. Commons 1, 67–87. Melack, J.M., Hess, L.L., 2010. Remote sensing of the distribution and extent of wetlands in the Amazon Basin. In: Junk, W.J., Piedade, M.T.F., Wittmann, F., Schöngart, J., Parolin, P. (Eds.), Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management. Springer Verlag, pp. 43–59. Metzger, J.P., 2000. Tree functional group richness and landscape structure in a Brazilian tropical fragmented landscape. Ecol. Appl. 10, 1147–1161. https://doi.org/10.1890/ 1051-0761(2000) 010[1147:TFGRAL]2.0.CO;2. Naeem, S., Bunker, D.E., Hector, A., Loreau, M., Perrings, C., 2009. Biodiversity, Ecosystem Functioning, and Human Wellbeing: An Ecological and Economic Perspective. Oxford University Press, USA. Naveh, Z., 2007. Landscape ecology and sustainability. Landsc. Ecol. 22, 1437–1440. https://doi.org/10.1007/s10980-007-9171-x. Nepstad, D., Schwartzman, S., Bamberger, B., Santilli, M., Ray, D., Schlesinger, P., Lefebvre, P., Alencar, A., Prinz, E., Fiske, G., Rolla, A., 2006. Inhibition of Amazon deforestation and fire by parks and indigenous lands. Conserv. Biol. 20, 65–73. https://doi.org/10.1111/j.1523-1739.2006.00351.x. Osborn, F., Goitia, W., Cabrera, M., Jaffé, K., 1999. Ants, plants and butterflies as diversity indicators: comparisons between strata at six forest sites in Venezuela. Stud. Neotrop. Fauna Environ. 34, 85–92. Pantoja, M.C., 2005. A várzea do Médio Amazonas e a sustentabilidade de um modo de 419