Assessment and monitoring of long-term forest cover changes in Odisha, India using remote sensing and GIS

Deforestation and fragmentation are important concerns in managing and conserving tropical forests and have global significance. In the Indian context, in the last one century, the forests have undergone significant changes due to several policies undertaken by government as well as increased population pressure. The present study has brought out spatiotemporal changes in forest cover and variation in forest type in the state of Odisha (Orissa), India, during the last 75 years period. The mapping for the period of 1924–1935, 1975, 1985, 1995 and 2010 indicates that the forest cover accounts for 81,785.6 km² (52.5 %), 56,661.1 km² (36.4 %), 51,642.3 km² (33.2 %), 49,773 km² (32 %) and 48,669.4 km² (31.3 %) of the study area, respectively. The study found the net forest cover decline as 40.5 % of the total forest and mean annual rate of deforestation as 0.69 %?year^?1 during 1935 to 2010. There is a decline in annual rate of deforestation during 1995 to 2010 which was estimated as 0.15 %. Forest type-wise quantitative loss of forest cover reveals large scale deforestation of dry deciduous forests. The landscape analysis shows that the number of forest patches (per 1,000) are 2.463 in 1935, 10.390 in 1975, 11.899 in 1985, 12.193 in 1995 and 15.102 in 2010, which indicates high anthropogenic pressure on the forests. The mean patch size (km²) of forest decreased from 33.2 in 1935 to 5.5 in 1975 and reached to 3.2 by 2010. The study demonstrated that monitoring of long term forest changes, quantitative loss of forest types and landscape metrics provides critical inputs for management of forest resources.     

A comparative assessment of land cover dynamics of three protected forest areas in tropical eastern Africa

Processes of deforestation, known to threaten tropical forest biodiversity, have not yet been studied sufficiently in East Africa. To shed light on the patterns and causes of human influences on protected forest ecosystems, comparisons of different study areas regarding land cover dynamics and potential drivers are needed. We analyze the development of land cover since the early 1970s for three protected East African rainforests and their surrounding farmlands and assess the relationship between the observed changes in the context of the protection status of the forests. Processing of Landsat satellite imagery of eight or seven time steps in regular intervals results in 12 land cover classes for the Kakamega–Nandi forests (Kenya) and Budongo Forest (Uganda) whereas ten are distinguished for Mabira Forest (Uganda). The overall classification accuracy assessed for the year 2001 or 2003 is similarly high for all three study areas (81% to 85%). The time series reveal that, despite their protection status, Kakamega–Nandi forests and Mabira Forest experienced major forest decrease, the first a continuous forest loss of 31% between 1972/1973 and 2001, the latter an abrupt loss of 24% in the late 1970s/early 1980s. For both forests, the temporally dense time series show short-term fluctuations in forest classes (e.g., areas of forest regrowth since the 1980s or exotic secondary bushland species from the 1990s onwards). Although selectively logged, Budongo Forest shows a much more stable forest cover extent. A visual overlay with population distribution for all three regions clearly indicates a relationship between forest loss and areas of high population density, suggesting population pressure as a main driver of deforestation. The revealed forest losses due to local and commercial exploitation further demonstrate that weak management impedes effective forest protection in East Africa.     

Assessment of large-scale deforestation of Nawarangpur district, Orissa, India: a remote sensing based study

The Nawarangpur district, Orissa, a tropical region with Sal mixed moist deciduous and Sal mixed dry deciduous forests, has been affected by extensive deforestation. The district was surveyed using Landsat MSS (1973), Landsat TM (1990) and IRS P6 LISS III (2004) satellite imagery. From 1973 to 1990, more than 888.6 km(2) of dense forest (rate of deforestation = 3.62) and from 1990 to 2004, 429.7 km(2) (rate of deforestation = 3.97) were found to have been deforested. The analysis of results identified the reduction in area of dense forest and increase of agricultural land, degraded areas of abandoned agricultural land and unproductive scrub. There is an urgent need for rational management of the remaining forest for it to be able to survive beyond next decades. From this study it can be concluded that temporal changes and the factors affecting these changes should be determined for sustainable management of natural resources.     

Geospatial assessment and monitoring of historical forest cover changes (1920–2012) in Nilgiri Biosphere Reserve,Western Ghats,India

Deforestation in the biosphere reserves, which are key Protected Areas has negative impacts on biodiversity, climate, carbon fluxes and livelihoods. Comprehensive study of deforestation in biosphere reserves is required to assess the impact of the management effectiveness. This article assesses the changes in forest cover in various zones and protected areas of Nilgiri Biosphere Reserve, the first declared biosphere reserve in India which forms part of Western Ghats-a global biodiversity hotspot. In this study, we have mapped the forests from earliest available topographical maps and multi-temporal satellite data spanning from 1920’s to 2012 period. Mapping of spatial extent of forest cover, vegetation types and land cover was carried out using visual interpretation technique. A grid cell of 1 km?×?1 km was generated for time series change analysis to understand the patterns in spatial distribution of forest cover (1920–1973–1989–1999–2006–2012). The total forest area of biosphere reserve was found to be 5,806.5 km² (93.8 % of total geographical area) in 1920. Overall loss of forest cover was estimated as 1,423.6 km² (24.5 % of the total forest) with reference to 1920. Among the six Protected Areas, annual deforestation rate of >0.5 was found in Wayanad wildlife sanctuary during 1920–1973. The deforestation in Nilgiri Biosphere Reserve is mainly attributed to conversion of forests to plantations and agriculture along with submergence due to construction of dams during 1920 to 1989. Grid wise analysis indicates that 851 grids have undergone large-scale negative changes of >75 ha of forest loss during 1920–1973 while, only 15 grids have shown >75 ha loss during 1973–1989. Annual net rate of deforestation for the period of 1920 to 1973 was calculated as 0.5 followed by 0.1 for 1973 to 1989. Our analysis shows that there was large-scale deforestation before the declaration of area as biosphere reserve in 1986; however, the deforestation has drastically reduced after the declaration due to high degree of protection, thus indicating the secure future of reserve in the long term under the current forest management practices. The present work will stand as the most up-to-date assessment on the forest cover of the Nilgiri Biosphere Reserve with immediate applications in monitoring and management of forest biodiversity.     

Hurricane Katrina-induced forest damage in relation to ecological factors at landscape scale

Forest stand stability to strong winds such as hurricanes has been found to be associated with a number of forest, soil and topography factors. In this study, through applying geographic information system (GIS) and logit regression, we assessed effects of forest characteristics and site conditions on pattern, severity and probability of Hurricane Katrina disturbance to forests in the Lower Pearl River Valley, USA. The factors included forest type, forest coverage, stand density, soil great group, elevation, slope, aspect, and stream buffer zone. Results showed that Hurricane Katrina damaged 60% of the total forested land in the region. The distribution and intensity of the hurricane disturbance varied across the landscape, with the bottomland hardwood forests on river floodplains most severely affected. All these factors had a variety of effects on vulnerability of the forests to the hurricane disturbance and thereby spatial patterns of the disturbance. Soil groups and stand factors including forest types, forest coverage and stand density contributed to 85% of accuracy in modeling the probability of the hurricane disturbance to forests in this region. Besides assessment of Katrina's damage, this study elucidates the great usefulness of remote sensing and GIS techniques combined with statistics modeling in assessment of large-scale risks of hurricane damage to coastal forests.     

Finding threatened forest areas in the central volcanic mountain range conservation area in Costa Rica

The degree at which tropical forests are exposed to human pressure is spatially dependent. Population density, proximity to roads, terrain slope, logging activities and land distribution projects are well known factors inducing deforestation and forest degradation in Latin America. Using expert knowledge to weight these threat factors and a Geographical Information System for spatial modeling, a multi-criteria analysis procedure is presented, that allows stratifying a study region in categories of deforestation threat. The procedure was implemented in the Central Volcanic Mountain Range Conservation Area (CVMRCA) in Costa Rica with the purpose of finding areas with a combination of physical and socioeconomic characteristics that is particularly predisposing to a high probability of deforestation. To validate the map, the CVMRCA was stratified in categories of deforestation risk, and the result was superposed to historical deforestation data of the period 1986–1996. The good correlation between risk category and historical deforestation (r = 0.91, p < 0.001) indicates that the map can be used as a decision support tool for defining priority areas for conservation action.     

Fire risk evaluation using multicriteria analysis—a case study

Forest fires are one of the major causes of ecological disturbance and environmental concerns in tropical deciduous forests of south India. In this study, we use fuzzy set theory integrated with decision-making algorithm in a Geographic Information Systems (GIS) framework to map forest fire risk. Fuzzy set theory implements classes or groupings of data with boundaries that are not sharply defined (i.e., fuzzy) and consists of a rule base, membership functions, and an inference procedure. We used satellite remote sensing datasets in conjunction with topographic, vegetation, climate, and socioeconomic datasets to infer the causative factors of fires. Spatial-level data on these biophysical and socioeconomic parameters have been aggregated at the district level and have been organized in a GIS framework. A participatory multicriteria decision-making approach involving Analytical Hierarchy Process has been designed to arrive at a decision matrix that identified the important causative factors of fires. These expert judgments were then integrated using spatial fuzzy decision-making algorithm to map the forest fire risk. Results from this study were quite useful in identifying potential “hotspots” of fire risk, where forest fire protection measures can be taken in advance. Further, this study also demonstrates the potential of multicriteria analysis integrated with GIS as an effective tool in assessing “where and when” forest fires will most likely occur.     

The Deforestation of Rural Areas in the Lower Congo Province

The Lower Congo is one of eleven provinces in the Democratic Republic of Congo, and is located southwest of Kinshasa Town Province. It has an area of approximately 53.947 km² with a population of 1 504 361 at an estimated 237 persons per km². The Province comprises five districts, including Lukaya and Cataracts where rural poverty is severe and the population struggle to make a living through agriculture and woodcutting. These activities result in excessive resource exploitation. The high demand for foodstuffs and the high consumption of wood (for energy, construction and export) in Kinshasa, the capital city of the Democratic Republic of Congo and the expanding towns of Matadi and Boma in the Lower Congo Province, are speeding the deforestation rate and unbalancing forest ecosystems. In addition there is the stress resulting from reduced josher (the rest period for agriculture ground), plus climate change and erosion. The phenomena that that we need to address in these two districts include deforestation, reduced josher, excessive agriculture, erosion, burning and climate change which taken together largely explain the current soil degradation. These areas are marked by excessive post deforestation savannah formation and extended areas of sandy soil, distributed throughout grass and shrub savannahs. This desertification, which is rampant in Lukaya and Cataracts, risks imprisoning the rural population in a vicious cycle of poverty if adequate solutions are not found.     

Quantification of the Impact of Land-Use Changes on Ecosystem Services: A Case Study in Pingbian County, China

Pingbian Miao Autonomous County is one of the poorest rural areas in China. Land-use changes, mainly driven by agricultural expansion and deforestation, may significantly impact ecosystem services and functions, but such effects are difficult to quantify. In the present study, Landsat image data were combined with the published coefficients about the world and China ecosystem to quantify land-use and ecosystem service changes in the mountainous area. A sensitivity analysis was employed to determine the effect of manipulating these coefficients on the estimated values. Our results show that during the past decades (from 1973 to 2004) forests and grasslands were converted into shrubland and cropland, respectively, resulting in a continuous decrease in ecosystem service (from 124.5 US$ × 10⁶ in 1973 to 100.4 US$ × 10⁶ in 2004). We found that the decrease of mixed forest in the study area was the largest contributor (i.e., 25.4 US$ × 10⁶) to the decline of the ecosystem service. Therefore we propose that future land-use policy should pay more attention to the crucial ecosystem functions of these forests (including tropical forest), and that it is necessary to balance the relationship between the livelihood of local farmers and environmental protection in order to maintain a healthy and stable ecosystem.     

An ecosystem report on the Panama Canal: monitoring the status of the forest communities and the watershed

In 1996, the Smithsonian Tropical Research Institute and the Republic of Panama's Environmental Authority, with support fromthe United States Agency for International Development, undertook a comprehensive program to monitor the ecosystem of the Panama Canal watershed. The goals were to establish baselineindicators for the integrity of forest communities and rivers. Based on satellite image classification and ground surveys, the2790 km² watershed had 1570 km² of forest in 1997, 1080 km² of which was in national parks and nature monuments. Most of the 490 km² of forest not currently in protected areas lies along the west bank of the Canal, and its managementstatus after the year 2000 turnover of the Canal from the U.S. to Panama remains uncertain. In forest plots designed to monitorforest diversity and change, a total of 963 woody plant specieswere identified and mapped. We estimate there are a total of 850–1000 woody species in forests of the Canal corridor. Forestsof the wetter upper reaches of the watershed are distinct in species composition from the Canal corridor, and have considerably higher diversity and many unknown species. Theseremote areas are extensively forested, poorly explored, and harbor an estimated 1400–2200 woody species. Vertebrate monitoring programs were also initiated, focusing on species threatened by hunting and forest fragmentation. Large mammals are heavily hunted in most forests of Canal corridor, and therewas clear evidence that mammal density is greatly reduced in hunted areas and that this affects seed predation and dispersal. The human population of the watershed was 113 000 in 1990, and grew by nearly 4% per year from 1980 to 1990. Much of this growth was in a small region of the watershed on the outskirts of Panama City, but even rural areas, including villages near and within national parks, grew by 2% per year. There is no sewage treatment in the watershed, and many towns have no trashcollection, thus streams near large towns are heavily polluted. Analyses of sediment loads in rivers throughout the watershed did not indicate that erosion has been increasing as a result ofdeforestation, rather, erosion seems to be driven largely by total rainfall and heavy rainfall events that cause landslides.Still, models suggest that large-scale deforestation would increase landslide frequency, and failure to detect increases inerosion could be due to the gradual deforestation rate and the short time period over which data are available. A study of runoff showed deforestation increased the amount of water fromrainfall that passed directly into streams. As a result, dry season flow was reduced in a deforested catchment relative to aforested one. Currently, the Panama Canal watershed has extensive forest areasand streams relatively unaffected by humans. But impacts of hunting and pollution near towns are clear, and the burgeoningpopulation will exacerbate these impacts in the next few decades.Changes in policies regarding forest protection and pollution control are necessary.     

Analysis of spatio-temporal land cover changes for hydrological impact assessment within the Nyando River Basin of Kenya

The spatio-temporal changes in the land cover states of the Nyando Basin were investigated for auxiliary hydrological impact assessment. The predominant land cover types whose conversions could influence the hydrological response of the region were selected. Six Landsat images for 1973, 1986, and 2000 were processed to discern the changes based on a methodology that employs a hybrid of supervised and unsupervised classification schemes. The accuracy of the classifications were assessed using reference datasets processed in a GIS with the help of ground-based information obtained through participatory mapping techniques. To assess the possible hydrological effect of the detected changes during storm events, a physically based lumped approach for infiltration loss estimation was employed within five selected sub-basins. The results obtained indicated that forests in the basin declined by 20% while agricultural fields expanded by 16% during the entire period of study. Apparent from the land cover conversion matrices was that the majority of the forest decline was a consequence of agricultural expansion. The model results revealed decreased infiltration amounts by between 6% and 15%. The headwater regions with the vast deforestation were noted to be more vulnerable to the land cover change effects. Despite the haphazard land use patterns and uncertainties related to poor data quality for environmental monitoring and assessment, the study exposed the vast degradation and hence the need for sustainable land use planning for enhanced catchment management purposes.     

Carbon storage and emissions offset potential in an African dry forest, the Arabuko-Sokoke Forest, Kenya

Concerns about rapid tropical deforestation, and its contribution to rising atmospheric concentrations of greenhouse gases, increase the importance of monitoring terrestrial carbon storage in changing landscapes. Emerging markets for carbon emission offsets may offer developing nations needed incentives for reforestation, rehabilitation, and avoided deforestation. However, relatively little empirical data exists regarding carbon storage in African tropical forests, particularly for those in arid or semi-arid regions. Kenya's 416 km(2) Arabuko-Sokoke Forest (ASF) is the largest remaining fragment of East African coastal dry forest and is considered a global biodiversity hotspot (Myers et al. 2000), but has been significantly altered by past commercial logging and ongoing extraction. Forest carbon storage for ASF was estimated using allometric equations for tree biomass, destructive techniques for litter and herbaceous vegetation biomass, and spectroscopy for soils. Satellite imagery was used to assess land cover changes from 1992 to 2004. Forest and thicket types (Cynometra webberi dominated, Brachystegia spiciformis dominated, and mixed species forest) had carbon densities ranging from 58 to 94 Mg C/ha. The ASF area supported a 2.8-3.0 Tg C carbon stock. Although total forested area in ASF did not change over the analyzed time period, ongoing disturbances, quantified by the basal area of cut tree stumps per sample plot, correlated with decreased carbon densities. Madunguni Forest, an adjoining forest patch, lost 86% of its forest cover and at least 76% of its terrestrial carbon stock in the time period. Improved management of wood harvesting in ASF and rehabilitation of Madunguni Forest could substantially increase terrestrial carbon sequestration in the region.     

Mega-Development Trends in the Amazon: Implications for Global Change

This paper describes four global-change phenomena that are having major impacts on Amazonian forests. The first is accelerating deforestation and logging. Despite recent government initiatives to slow forest loss, deforestation rates in Brazilian Amazonia have increased from 1.1 million ha yr^–1 in the early 1990s, to nearly 1.5 million ha yr^–1 from 1992–1994, and to more than 1.9 million ha yr^–1 from 1995–1998. Deforestation is also occurring rapidly in some other parts of the Amazon Basin, such as in Bolivia and Ecuador, while industrialized logging is increasing dramatically in the Guianas and central Amazonia.The second phenomenon is that patterns of forest loss and fragmentation are rapidly changing. In recent decades, large-scale deforestation has mainly occurred in the southern and eastern portions of the Amazon — in the Brazilian states of Pará, Maranho, Rondônia, Acre, and Mato Grosso, and in northern Bolivia. While rates of forest loss remain very high in these areas, the development of major new highways is providing direct conduits into the heart of the Amazon. If future trends follow past patterns, land-hungry settlers and loggers may largely bisect the forests of the Amazon Basin.The third phenomenon is that climatic variability is interacting with human land uses, creating additional impacts on forest ecosystems. The 1997/98 El Niño drought, for example, led to a major increase in forest burning, with wildfires raging out of control in the northern Amazonian state of Roraima and other locations. Logging operations, which create labyrinths of roads and tracks in forsts, are increasing fuel loads, desiccation and ignition sources in forest interiors. Forest fragmentation also increases fire susceptibility by creating dry, fire-prone forest edges.Finally, recent evidence suggests that intact Amazonian forests are a globally significant carbon sink, quite possibly caused by higher forest growth rates in response to increasing atmospheric CO₂ fertilization. Evidence for a carbon sink comes from long-term forest mensuration plots, from whole-forest studies of carbon flux and from investigations of atmospheric CO₂ and oxygen isotopes. Unfortunately, intact Amazonian forests are rapidly diminishing. Hence, not only is the destruction of these forests a major source of greenhouse gases, but it is reducing their intrinsic capacity to help buffer the rapid anthropogenic rise in CO₂.     

Incentive mechanisms to enforce sustainable forest exploitation

Using a two-player differential game approach, this paper deals with the issue of tropical deforestation. The assumption is that developing forestry countries (called the South) do not, or cannot, necessarily have the same utility for forest conservation than the developed countries (called the North). Given the global nature of the problem, we assume that it may be wise that the North offers financial support to the South to help in achieving a sustainable forest policy. We derive sustainable and short-run deforestation policies and provide a transfer mechanism which insures that the South will indeed implement the sustainable one.     

Regional impact analysis of climate change on natural and managed forests in the Federal State of Brandenburg, Germany

A methodology for regional application of forest simulation models has been developed as part of an assessment of possible climate change impacts in the Federal state of Brandenburg (Germany). Here we report on the application of a forest gap model to analyse the impacts of climate change on species composition and productivity of natural and managed forests in Brandenburg using a statistical method for the development of climate scenarios. The forest model was linked to a GIS that includes soil and groundwater table maps, as well as gridded climate data with a resolution of 10 × 10 km and simulated a steady-state species composition which was classified into forest types based on the biomass distribution between species. Different climate scenarios were used to assess the sensitivity of species composition to climate change. The simulated forest distribution patterns for current climate were compared with a map of Potential Natural Vegetation (PNV) of Brandenburg.In order to analyse the possible consequences of climate change on forest management, we used forest inventory data to initialize the model with representative forest stands. Simulation experiments with two different management strategies indicated how forest management could respond to the projected impacts of climate change. The combination of regional analysis of natural forest dynamics under climate change with simulation experiments for managed forests outlines possible trends for the forest resources. The implications of the results are discussed, emphasizing the regional differences in environmental risks and the adaptation potentials of forestry in Brandenburg.     

Use of forest inventories and geographic information systems to estimate biomass density of tropical forests: Application to tropical Africa

One of the most important databases needed for estimating emissions of carbon dioxide resulting from changes in the cover, use, and management of tropical forests is the total quantity of biomass per unit area, referred to as biomass density. Forest inventories have been shown to be valuable sources of data for estimating biomass density, but inventories for the tropics are few in number and their quality is poor. This lack of reliable data has been overcome by use of a promising approach that produces geographically referenced estimates by modeling in a geographic information system (GIS). This approach has been used to produce geographically referenced, spatial distributions of potential and actual (circa 1980) aboveground biomass density of all forests types in tropical Africa. Potential and actual biomass density estimates ranged from 33 to 412 Mg ha^–1 (10⁶g ha^–1) and 20 to 299 Mg ha^–1, respectively, for very dry lowland to moist lowland forests and from 78 to 197 Mg ha^–1 and 37 to 105 Mg ha^–1, respectively, for montane-seasonal to montane-moist forests. Of the 37 countries included in this study, more than half (51%) contained forests that had less than 60% of their potential biomass. Actual biomass density for forest vegetation was lowest in Botswana, Niger, Somalia, and Zimbabwe (about 10 to 15 Mg ha^–1). Highest estimates for actual biomass density were found in Congo, Equatorial Guinea, Gabon, and Liberia (305 to 344 Mg ha^–1). Results from this research effort can contribute to reducing uncertainty in the inventory of country-level emission by providing consistent estimates of biomass density at subnational scales that can be used with other similarly scaled databases on change in land cover and use.     

Carbon Sequestration Potential of Indian Forests

The forestry sector can not only sustain its carbon but also has the potential to absorb carbon from the atmosphere. India has maintained approximately 64 Mha of forest cover for the last decade. The rate of afforestation in India is one of the highest among the tropical countries, currently estimated to be 2 Mha per annum. The annual productivity has increased from 0.7 m³ per hactare in 1985 to 1.37 m³ per hectare in 1995. Increase in annual productivity directly indicates an increase in forest biomass and hence higher carbon sequestration potential. The carbon pool for the Indian forests is estimated to be 2026.72 Mt for the year 1995. Estimates of annual carbon uptake increment suggest that our forests and plantations have been able to remove at least 0.125 Gt of CO₂ from the atmosphere in the year 1995. Assuming that the present forest cover in India will sustain itself with a marginal annual increase by 0.5 Mha in area of plantations, we can expect our forests to continue to act as a net carbon sink in future.     

Spatial Distribution and Temporal Change of Carbon Storage in Timber Biomass of Two Different Forest Management Units

Forests make up large ecosystems and can play an important role in mitigating the emissions of CO₂, the most important greenhouse gas. However, they are sources of atmospheric carbon when they are disturbed by human or natural causes. Storage of carbon through expansion and adaptive management of forest ecosystems can assist in reducing carbon concentrations in atmosphere. This study proposes a methodology to produce spatially explicit estimates of the carbon storages (aboveground plus belowground) depending on land use/cover changes in two different forest ecosystems during various periods. Carbon storages for each forest ecosystem were projected according to inventory data, and carbon storages were mapped in a geographic information system (GIS). Results showed that total carbon stored in above and belowground of both forest ecosystems increased from one period to other because of especially increase of productive forest areas and decline of degraded forest areas as well as protection of spruce forests subject to insect attacks.     

MAPPING TROPICAL DEFORESTATION IN CENTRAL AFRICA

The NASA Landsat Pathfinder Humid Tropical Deforestation Project was to map deforestation activities in the humid tropics using datasets from both the Landsat TM (Thematic Mapper) and MSS (Multispectral Scanner System). In Central Africa, its effort had been constrained by the availability of cloud-free satellite coverage, especially for the 1970s Landsat MSS imagery. Here, we reported the deforestation rate and its spatial variability in the region using 18 pairs of co-registered Landsat TM imagery from the 1980s to 1990s. Of the total classified area of 416000 km, there were approximately 217000 km² of dense forest and 24000 km² of degraded forest in the 1980s. A total of 1012 km² of forest, including 542 km² of dense forest and 470 km² of degraded forest, were cleared annually with an annual deforestation rate of 0.42%, varying among scenes ranging from 0.03 to 2.72%. Additionally, an average of 0.12% (ranging from 0.01 to 0.77% among scenes) or 257 km² of dense forest was degraded annually. Regression analyses indicated that extensive deforestation occurred in areas with larger forest cover, including dense and degraded forests. Image interpretation also confirmed the hypothesized relationship between deforestation and forest accessibility. The annual clearance of the dense forest was significantly related to the rural population density, and there was a positive relationship between the dense forest degraded during the 1980s–1990s and the degraded forest area in the 1980s.     

A framework for assessment and monitoring of small mammals in a lowland tropical forest

Development projects in tropical forests can impact biodiversity.Assessment and monitoring programs based on the principles of adaptive management assist managers to identify and reduce suchimpacts. The small mammal community is one important component ofa forest ecosystem that may be impacted by development projects. In 1996, a natural gas exploration project was initiated in a Peruvian rainforest. The Smithsonian Institution's Monitoring andAssessment of Biodiversity program cooperated with Shell Prospecting and Development Peru to establish an adaptive management program to protect the region's biodiversity. In thisarticle, we discuss the role of assessing and monitoring small mammals in relation to the natural gas project. We outline theconceptual issues involved in establishing an assessment andmonitoring program, including setting objectives, evaluating the results and making appropriate decisions. We also summarizethe steps taken to implement the small mammal assessment, provideresults from the assessment and discuss protocols to identifyappropriate species for monitoring.