Classification Of Landsat Images Research Articles

Developments in Landsat Land Cover Classification Methods: A Review

Land cover classification of Landsat images is one of the most important applications developed from Earth observation satellites. The last four decades were marked by different developments in land cover classification methods of Landsat images. This paper reviews the developments in land cover classification methods for Landsat images from the 1970s to date and highlights key ways to optimize analysis of Landsat images in order to attain the desired results. This review suggests that the development of land cover classification methods grew alongside the launches of a new series of Landsat sensors and advancements in computer science. Most classification methods were initially developed in the 1970s and 1980s; however, many advancements in specific classifiers and algorithms have occurred in the last decade. The first methods of land cover classification to be applied to Landsat images were visual analyses in the early 1970s, followed by unsupervised and supervised pixel-based classification methods using maximum likelihood, K-means and Iterative Self-Organizing Data Analysis Technique (ISODAT) classifiers. After 1980, other methods such as sub-pixel, knowledge-based, contextual-based, object-based image analysis (OBIA) and hybrid approaches became common in land cover classification. Attaining the best classification results with Landsat images demands particular attention to the specifications of each classification method such as selecting the right training samples, choosing the appropriate segmentation scale for OBIA, pre-processing calibration, choosing the right classifier and using suitable Landsat images. All these classification methods applied on Landsat images have strengths and limitations. Most studies have reported the superior performance of OBIA on different landscapes such as agricultural areas, forests, urban settlements and wetlands; however, OBIA has challenges such as selecting the optimal segmentation scale, which can result in over or under segmentation, and the low spatial resolution of Landsat images. Other classification methods have the potential to produce accurate classification results when appropriate procedures are followed. More research is needed on the application of hybrid classifiers as they are considered more complex methods for land cover classification.

Ontology-based classification of remote sensing images using spectral rules

Earth Observation data is of great interest for a wide spectrum of scientific domain applications. An enhanced access to remote sensing images for “domain” experts thus represents a great advance since it allows users to interpret remote sensing images based on their domain expert knowledge. However, such an advantage can also turn into a major limitation if this knowledge is not formalized, and thus is difficult for it to be shared with and understood by other users. In this context, knowledge representation techniques such as ontologies should play a major role in the future of remote sensing applications. We implemented an ontology-based prototype to automatically classify Landsat images based on explicit spectral rules. The ontology is designed in a very modular way in order to achieve a generic and versatile representation of concepts we think of utmost importance in remote sensing. The prototype was tested on four subsets of Landsat images and the results confirmed the potential of ontologies to formalize expert knowledge and classify remote sensing images.

Analyzing Parcel-Level Relationships between Urban Land Expansion and Activity Changes by Integrating Landsat and Nighttime Light Data

Urban growth is a process that imposes profound physical and socioeconomic restructuring on cities. Urban land expansion as an immediate physical manifestation of urban growth has been extensively studied using a variety of remote sensing methods. However, little research addresses the interactions between urban land expansion and corresponding activity changes, especially at local scales. We propose an innovative analytical framework that integrates Landsat and nighttime light data to capture the parcel-level relationships between urban land expansion and activity changes. The urban land data are acquired based on the classification of Landsat images, whereas the activity changes are approximated by the nighttime light data. Using the Local Indicator of Spatial Association (LISA) (local Moran’s I) approach, four types of local relationships between land expansion and activity changes are defined at the parcel level. The proposed analytical framework is applied in Guangzhou, China, as a case study. The results reveal the mismatched growth between urban land and activity intensity at the parcel level, where the increase in urban land area outpaces the increase of activity intensity. Such results are expected to provide a more comprehensive understanding of urban growth, and can be used to assist urban planning and related decision-making.

Assessment of multifunctional landscapes dynamics in the mountainous basin of the Mo River (Togo, West Africa)

In this study, historical landscape dynamics were investigated to (i) map the land use/cover types for the years 1972, 1987, 2000 and 2014; (ii) determine the types and processes of landscape dynamics; and (iii) assess the landscape fragmentation and habitat loss over time. Supervised classification of multi-temporal Landsat images was used through a pixel-based approach. Post–classification methods included systematic and random change detection, trajectories analysis and landscape fragmentation assessment. The overall accuracies (and Kappa statistics) were of 68.86% (0.63), 91.32% (0.79), 90.66% (0.88) and 91.88% (0.89) for 1972, 1987, 2000 and 2014, respectively. The spatio-temporal analyses indicated that forests, woodlands and savannahs dominated the landscapes during the four dates, though constant areal decreases were observed. The most important dynamic process was the decline of woodlands with an average annual net loss rate of–2%. Meanwhile, the most important land transformation occurred during the transition 2000–2014, due to anthropogenic pressures. Though the most important loss of vegetation greenness occurred in the unprotected areas, the overall analyses of change indicated a declining trend of land cover quality and an increasing landscape fragmentation. Sustainable conservation strategies should be promoted while focusing restoration attention on degraded lands and fragmented ecosystems in order to support rural livelihood and biodiversity conservation.

Mô phỏng sự thay đổi diện tích đất lúa dưới tác động của hạn, mặn - Trường hợp nghiên cứu ở tỉnh Sóc Trăng

In recent years, drought and salinity intrusion frequently affect rice cultivation areas in dry season of coastal provinces in the Mekong Delta, especially Soc Trang province. This paper aims to analyze the changes of rice areas in dry season by classifying Landsat images based on decision tree with the normalized difference vegetation index data and to build a simulation model of rice area changes under the impacts of drought and salinity intrusion. The rice maps of Soc Trang in dry season in 2014 and in 2016 are developed, corresponding to the maps in normal temperature and in drought and salinity intrusion conditions. These classified maps are assessed by Kappa coefficient of 0.89 for the map in 2014 and 0.83 for the one in 2016. Then, the model of rice area changes is built based on rainfall data, salinity intrusion into irrigation regions. The simulated result of rice map of dry season in 2016 is compared with the real one with Kappa value of 0.88. This result shows the simulated map can be used to provide a visual assessment of rice area under drought and salinity conditions.

Conservation Reserve Program mitigates grassland loss in the lesser prairie-chicken range of Kansas

Since the beginning of the 20th century, the overall occupied range of the lesser prairie-chicken (Tympanuchus pallidicinctus) has declined by 84% commensurate with population trends. Much of this decline has been attributed to the loss and fragmentation of native grasslands throughout the lesser prairie-chicken range. However, quantification of changes in land cover in the distribution of the lesser prairie-chicken is lacking. Our objectives were to (1) document changes in the areal extent and connectivity of grasslands in the identified lesser prairie-chicken range in Kansas, USA, (>60% of extant lesser prairie-chicken population) from the 1950s to 2013 using remotely sensed data and (2) assess the potential of the Conservation Reserve Program (U.S. Department of Agriculture Program converting cropland to permanent cover; CRP) to mitigate grassland loss. Digital land cover maps were generated on a decadal time step through spectral classification of LANDSAT images and visual analysis of aerial photographs (1950s and 1960s). Landscape composition and configuration were assessed using FRAGSTATS to compute a variety of landscape metrics measuring changes in the amount of grassland present as well as changes in the size and configuration of grassland patches. With the exception of a single regional portion of the range, nearly all of the grassland converted to cropland in the lesser prairie-chicken range of Kansas occurred prior to the 1950s. Prior to the implementation of CRP, the amount of grassland decreased 3.6% between the 1950s and 1985 from 18,455 km2 to 17,788 km2. Since 1985, the overall amount of grassland in the lesser prairie-chicken range has increased 11.9% to 19,898 km2 due to implementation of CRP, although the area of grassland decreased between 1994 and 2013 as CRP contracts were not renewed by landowners. Since 1986 grassland in Kansas became more connected and less fragmented in response to the CRP. While the CRP has been successful in increasing grassland quantity and connectivity throughout the lesser prairie-chicken range in Kansas, offsetting loss of grassland since the 1950s, abundance and occupied range of lesser prairie-chickens has declined since the 1980s, suggesting that habitat quality is the principal factor influencing population demography of the species. Although the CRP is contributing to conservation actions for lesser prairie-chickens, efforts to improve habitat quality throughout the range of the lesser prairie-chicken are likely necessary to meet management goals. Continuation of the CRP faces an uncertain future in the face of rising commodity prices, energy development, and reduction in program scope, leaving open the possibility that these areas that have created habitat for lesser prairie-chickens could be lost.

A Spectral Signature Shape-Based Algorithm for Landsat Image Classification

Land-cover datasets are crucial for earth system modeling and human-nature interaction research at local, regional and global scales. They can be obtained from remotely sensed data using image classification methods. However, in processes of image classification, spectral values have received considerable attention for most classification methods, while the spectral curve shape has seldom been used because it is difficult to be quantified. This study presents a classification method based on the observation that the spectral curve is composed of segments and certain extreme values. The presented classification method quantifies the spectral curve shape and takes full use of the spectral shape differences among land covers to classify remotely sensed images. Using this method, classification maps from TM (Thematic mapper) data were obtained with an overall accuracy of 0.834 and 0.854 for two respective test areas. The approach presented in this paper, which differs from previous image classification methods that were mostly concerned with spectral “value” similarity characteristics, emphasizes the "shape" similarity characteristics of the spectral curve. Moreover, this study will be helpful for classification research on hyperspectral and multi-temporal images.

Monitoring urban growth by using segmentation-classification of multispectral Landsat images in Izmit, Turkey.

Assessing the spatial land use and land cover (LULC) information is essential for decision making and management of landscapes. In fact, LULC information has been changed dramatically in fast-growing cities. This results in wrong land use problems due to unplanned and uncontrolled urbanization. The planning and evaluating of limited natural resources under the pressure of a growing population can be possible when a precise land use management plan is established. Therefore, it is imperative to monitor continuous LULC changes for future planning. Remote sensing (RS) technique is used for determining changes in LULC in urban areas. In this study, we have focused on Izmit, which is one of a growing number of metropolitan cities where the impact of the spatial growing period on LULC has been assessed over the past 30years by using RS data. We have utilized the segmentation process and supervised classification of Landsat satellite images for four different dates (1985, 1995, 2005, and 2015). The outcome of this research can be summarized by significant changes in the shares of urban areas and farmland LULC classes. The overall observed increase in urban area class is up to 2177ha between 1985 and 2015 period and this dramatic change has resulted in the decline of 1211ha of farmland. Another conclusion is that the new residential areas have been created to the north, south and east of Izmit during this period.

Assessing the Impacts of Land Use and land cover change on Pastoral Livestock Farming in South-Eastern Burkina Faso

<p class="1Body">Conversion of pastures to cropland is one of the most important issues facing livestock farming in Burkina Faso. This study examined the impact of land use/cover change on pastoral livestock farming in Boulgou province between 1980 and 2013. Landsat satellite images (1989, 2001 and 2013) and socio-economic data were analysed. The interpretation of the classified Landsat images revealed an increase in cropland from 20.5% in 1989 to 36.7% in 2013. This resulted mainly from the conversion of woody savannah and shrub and grass savannah to cropland. Pastoral livestock farmers reported that the major drivers of vegetation loss were drought (95.1 %), population growth (91.8%), cropland increase (91.4%), extraction of fuel wood (69.8%) and increase in livestock population (65.4). These changes affect livestock farming through reduction of pasture, poor access to water and reduction of livestock mobility routes according to the farmers. This calls for regional and national policies to protect grazing areas in Burkina Faso that are similar to policies being implemented for forest and other types of vegetation cover in other countries. For such pastoral policies to be successful, issues concerning the mobility of livestock farmers must be enshrined into such policies and this study is an example of information source for these policies.</p>

Mapping and Hydrologic Attribution of Temporary Wetlands Using Recurrent Landsat Imagery

In areas with a high density of ephemeral wetlands, traditional mapping protocols may underestimate occurrence of wetlands when single-date base-imagery is utilized. In the Pleistocene Sand Dunes Ecoregion in Oklahoma, National Wetland Inventory (NWI) maps created using base-imagery from a dry year omitted large numbers of ephemeral wetlands. To improve the likelihood of capturing inundated depressions, we classified water pixels from 51 Landsat images (3 images per year: pre/early, peak, and late/post growing season) from 1994 to 2011. Several image classification methods were tested but decision tree analysis with training pixels from multi-season imagery provided the greatest accuracy. Accuracy was determined through manual comparison of two Landsat images with concurrent aerial imagery (Kappa =0.96 and 0.93 for the two images). Wetland polygons were created from water/non-water rasters and given hydroperiod designations based on the number of inundated periods. Landsat-derived wetland maps identified 3156 wetland units, 718 more than the original 1980s NWI, with only 33.9 % agreement between the two maps. Finally, one meter LiDAR data were combined with classified Landsat images to determine the volume of water in wetlands during each image date. These wetland maps can assist with estimating the availability of inundated habitat during wet, dry, and average rainfall periods.

What are the Probabilities of Land-Use Transitions? The Answer Depends on the Classification Method

GIScience 2016 Short Paper Proceedings What are the Probabilities of Land-Use Transitions? The Answer Depends on the Classification Method Yulia Grinblat 1, 2 , Michael Gilichinsky 3 and Itzhak Benenson 1 Department of Geography and Human Environment, Tel Aviv University, Tel-Aviv Email: juliagri@post.tau.ac.il, bennya@post.tau.ac.il The Porter School of Environmental Studies, Tel Aviv University, Tel-Aviv Elbit Systems, Rehovot Email: gilichinsky@gmail.com Abstract Based on four time intervals within a 36-year period, we construct Land-Use/Land- Cover (LULC) maps and estimate the transition probabilities between six LULC states: built-up, agriculture, green and open spaces, transportation, and water surfaces. The LULC maps and transition probabilities matrices (TPM) were built based on the manual classification of high-resolution aerial photos and multispectral Landsat images for the same years. We considered the maps and TPM constructed from the aerial photos as a control, and compared them to those constructed from the Landsat images classified with several methods: mean-shift segmentation followed by Random Forest classification methods, and three pixel-based methods of classification: K-means, ISODATA, and maximum likelihood. For each classification the TPM were compared to the TPM constructed from the aerial photos. The goodness of fit of all maps obtained with the pixel-based methods was insufficient for estimating the LULC TPM. The LULC map obtained with the object- based classification method fit well to that based on the aerial photos, but the estimates of TMP were qualitatively different from those constructed from the aerial photos. This article raises doubts regarding the adequacy of Landsat data and standard classification methods for establishing LULC CA model rules, and calls for the careful reexamination of the entire land-use CA framework. 1. Introduction Conceptual simplicity and the ability of explicit representation of landscapes and their changes make Cellular Automata (CA) a standard tool for simulating urban and regional land-use dynamics. Typically, the CA models focus on estimating the rules of the LULC changes and analysis of the simulation results. However, the models put aside the uncertainty of the LULC maps that are used for establishing the transition rules. The major source of data for the CA modeling is Remote Sensing (RS) multispectral imagery classified for establishing LULC dynamics. It is often reported that the CA models are quite successful in predicting LULC, with the high overall fit (80-90%) between the real LULC and model outputs. This is indeed true when the validation is based on comparing the entire modeled area. However, as far as initial area is excluded from the comparison, the spatial fit between the predicted and real changes drops down (Hagen-Zanker et al. 2005; Pontius and Petrova 2010). A hierarchy of reasons of limited capacity of the CA models for predicting LULC changes can be proposed: (1) CA framework as a whole is insufficient for predict

Spatiotemporal Mining of Time-Series Remote Sensing Images Based on Sequential Pattern Mining

Abstract. With the continuous development of satellite techniques, it is now possible to acquire a regular series of images concerning a given geographical zone with both high accuracy and low cost. Research on how best to effectively process huge volumes of observational data obtained on different dates for a specific geographical zone, and to exploit the valuable information regarding land cover contained in these images has received increasing interest from the remote sensing community. In contrast to traditional land cover change measures using pair-wise comparisons that emphasize the compositional or configurational changes between dates, this research focuses on the analysis of the temporal sequence of land cover dynamics, which refers to the succession of land cover types for a given area over more than two observational periods. Using a time series of classified Landsat images, ranging from 2006 to 2011, a sequential pattern mining method was extended to this spatiotemporal context to extract sets of connected pixels sharing similar temporal evolutions. The resultant sequential patterns could be selected (or not) based on the range of support values. These selected patterns were used to explore the spatial compositions and temporal evolutions of land cover change within the study region. Experimental results showed that continuous patterns that represent consistent land cover over time appeared as quite homogeneous zones, which agreed with our domain knowledge. Discontinuous patterns that represent land cover change trajectories were dominated by the transition from vegetation to bare land, especially during 2009–2010. This approach quantified land cover changes in terms of the percentage area affected and mapped the spatial distribution of these changes. Sequential pattern mining has been used for string mining or itemset mining in transactions analysis. The expected novel significance of this study is the generalization of the application of the sequential pattern mining method for capturing the spatial variability of landscape patterns, and their trajectories of change, to reveal information regarding process regularities with satellite imagery.

Interfacing the geographic information system, remote sensing, and the soil conservation service–curve number method to estimate curve number and runoff volume in the Asir region of Saudi Arabia

Runoff coefficients are widely used as a diagnostic variable of runoff generation in process studies and as an important input parameter in hydrologic design. Many regions of Saudi Arabia do not have sufficient historical records and the detailed runoff information needed for physically based distribution models. The US Department of Agriculture, Natural Resources Conservation Service Curve Number (USDA-NRCS-CN) method was used in this study for determining the curve number and runoff depth for Asir region, Saudi Arabia. Runoff curve number was determined based on the factors of hydrologic soil group, land use and land cover, and slope gradient in the Asir region of Saudi Arabia using the geographic information system (GIS). The region was not subjected before for any previous studies concerning runoff generation or rainwater harvesting by using GIS and remote sensing. ArcGIS 9.3 and the global positioning system were used to develop the soil map on the basis of samples collected during a field survey. Soil textures were identified and used to build a map of the hydrological soil groups. Unsupervised and supervised classification of Landsat images was used to generate land use and land cover maps. The land cover distribution in the study area revealed that the largest class in the area, barren and sparsely vegetated land, occupied 62,832.38 km2 of the study area (more than 80 % of the total area), followed by shrubland (15,212.66 km2), cropland and pasture (52.34 km2), mixed shrubland/grassland (40.89 km2), and grassland (12.27 km2). Built-up land occupied only 11.45 km2. The slope map for the Asir region was generated from a 15-m digital elevation model. The GIS technique was used to develop the CN for the region based on the Soil Conservation Service method. The annual runoff depth was derived from the annual rainfall and CN per pixel using the raster calculator tool in ArcGIS. The rainfall distribution in the study area showed the wise identification of suitable sites for rainwater harvesting, as most of the constructed dams were located in the areas with higher rainfall. The analysis also revealed that the annual runoff for the study area ranged from 27 to 69 % of the total rainfall, with variation as low as 80 to 300 mm/year. This is the highest amount of runoff that can be generated in Saudi Arabia. The runoff volume was calculated for the entire region using the previously developed runoff depth map and basin areas. The results showed that the runoff volume (M3) in the Asir region varied from as low as 237,000 m3 to a maximum of 2,140,000 m3; this result revealed a significant yearly amount of runoff that can be harvested for any use. Moreover, the calculated runoff depth in the area agreed with the actual dam capacities for the existing dams in the Asir region. This approach can be applied in other regions in Saudi Arabia for rainwater harvesting and groundwater recharge.

Mapping and spatial characterization of Argentine High Andean peatbogs

Peatbogs are the most productive and biodiversity rich ecosystems of high dry Puna region, and provide essential ecosystems services to local inhabitants. Despite their ecological and economic importance, the geographic patterns and distribution of peatbogs as well the dynamic of this system are barely known. In this work we (1) identified and mapped subtropical Argentine High Andean peatbogs based in a supervised classification of Landsat images, and a posteriori spatial processing to define functional units; (2) used such classification to characterize the whole Argentina Puna region (14.3 million hectares) in terms of their geographic patterns of peatbogs; and (3) characterized the region’s sub-watersheds according to their peatbogs density and geographic characteristics. The post-process map reports a total area of peatbogs of 94427.55 ha (0.66 % of total study area) with 10,428 polygons. The majority of peatbogs are small or median (i.e., 21.8 % of total area included 10,000 ha represent 18 % of total area. Peatbogs density is spatially heterogeneous, with much higher density in the north and central east of the study area. Sub-watershed can be grouped into six main groups according to the percentage of peatbogs cover, mean size of peatbogs and altitude. The combination of basic information, such as map of peatbogs cover with spatial patterns characterization, is a priority input for the conservation planning of this extensive and valuable ecosystem, and to ongoing land use planning initiatives.

Examining Urban Expansion in the Greater Toronto Area Using Landsat Imagery from 1974–2014

The Greater Toronto Area is the most vital economic centre in Canada and has experienced rapid urban expansion in the past 40 years. This research uses Landsat images to detect the spatial and temporal dynamics of urban expansion in the Greater Toronto Area from 1974 to 2014. We quantitatively analyzed the extent of urban expansion and spatial patterns of growth from classified Landsat images. We then integrated our expansion findings with population data to observe the relationships between urban growth and population. We found that the Greater Toronto Area had significant growth of 1115 km2, expanding mainly in radiated and ribbon expansion modes. There was substantial correlation between urban extent and population in the period of study. These results demonstrate the efficacy of combining statistical population data with remote sensing imagery for the analysis of urban expansion.

Cambios en la cobertura arbolada de comunidades indígenas con y sin iniciativas de conservación, en Oaxaca, México

We analyzed the dynamics of forest cover change between 1990-2010 in eleven Chinantec indigenous communities in the Sierra Norte of the state of Oaxaca, Mexico. Five communities exhibited collective action for forest conservation (with-cac) and the remaining six did not (without-cac). Using supervised classification of Landsat satellite images from three dates (1990, 2000 and 2010) we obtained maps of forested/non-forested conditions. We then crossed maps for 1990-2000 and 2000-2010 to generate maps of change processes (deforestation and re-vegetation) and annual rates of change. In general, forest cover was relatively stable (~90%), but there were significant differences between the forest cover in communities with-cac and without-cac. In the first, there was greater forest cover, less permanence of the non-forested condition, and lower rates of change. In the communities without-cac there was greater re-vegetation between 2000-2010, but forest recovery does not appear to have been intentional, since the communities had no initiatives for taking care of their forests. In the communities with-cac there has been influence of advisors, academics, and programs with governmental and non-governmental entities, and these have had little presence in the without-cac communities. Improved understanding of the dynamics of high ecological value forests in common property, such as those in the northern part of Oaxaca State, is fundamental for better orientation of public policy focused at local forest in Mexico.

Assessing land-use changes driven by river dynamics in chronically flood affected Upper Brahmaputra plains, India, using RS-GIS techniques

This work documents land-use changes driven by river dynamics along two tributaries in the chronically flood affected Upper Brahmaputra floodplain which supports a population of more than half a million. Planform changes for a period of 40years are documented using topographical map and Landsat data, and the associated land-use change is assessed by utilising hybrid classification in GIS environment. Quantification of bankline migration shows that the river courses are unstable. A reversal in the rate of erosion and deposition is also observed. Hybrid classification of Landsat images yielded a higher level of accuracy as evident from the confusion matrixes. Overall, the accuracy of land-use classification ranged between 88.5% and 96.25%. Land-use change shows that there is an increase in settlement and agriculture and a decrease in the grassland. The area affected by erosion–deposition and river migration comprises primarily of the agricultural land. Effect of river dynamics on settlements is also evident. Loss of agricultural land and homestead led to the loss of livelihood and internal migration in the floodplains. The observed pattern of river dynamics and the consequent land-use change in the recent decades have thrown newer environmental challenges at a pace and magnitude way beyond the coping capabilities of the dwellers.

Estimating the Contribution of Loose Deposits to Potential Landslides Over Wenchuan Earthquake Zone, China

Loose deposits may lead to catastrophic landslides and rockslides especially during or after heavy rainfall events. In this paper, we propose a method to estimate the volume and spatial distribution (VSD) of loose deposits using airborne and spaceborne optical remote sensing data and Interferometric Synthetic Aperture Radar (InSAR) coherence. The estimated VSD of coseismic loose deposits in Wenchuan, China, the epicenter of the Sichuan Earthquake, is then used to predict potential landslides. First, high-resolution digital elevation models (DEMs) and digital orthophoto map (DOM) are developed using stereo imagery collected by an unmanned aerial vehicle (UAV) over a core test site. Next, the VSD of loose deposits is determined using the UAV imagery, DEM, and a landslide inventory map developed by field surveys, and a power–law relationship between the area and volume is established. In order to determine the loose deposits of the entire study area, the InSAR coherence of PALSAR images collected before and right after the earthquake, Landsat image classification, and DEM combined with Analytical Hierarchy Process (AHP) model are integrated in a decision tree algorithm. Finally, we develop a quantitative method of determining the severity and likelihood of potential geohazards based on VSD of coseismic loose deposits and other variables including the Normalized Difference Vegetation Index (NDVI), slope gradient, elevation, and land-cover/land-use (LCLU). A total of 42 catchments area are identified as the hot spots with a very high or high possibility of potential landslides. The VSD of coseismic loose deposits over these catchments is about 92.34% and 77.96% of the total potential loose deposits area.

Decadal monitoring of the Niger Inner Delta flood dynamics using MODIS optical data

Wetlands provide a vital resource to ecosystem services and associated rural livelihoods but their extent, geomorphological heterogeneity and flat topography make the representation of their hydrological functioning complex. A semi automated method exploiting 526 MODIS (Moderate Resolution Imaging Spectroradiometer) 8-day 500 m resolution images was developed to study the spatial and temporal dynamics of the annual flood across the Niger Inner Delta over the period 2000–2011. A composite band ratio index exploiting the Modified Normalised Difference Water Index (MNDWI) and Normalised Difference Moisture Index (NDMI) with fixed thresholds provided the most accurate detection of flooded areas out of six commonly used band ratio indices. K-means classified Landsat images were used to calibrate the thresholds. Estimated flooded surface areas were evaluated against additional classified Landsat images, previous studies and field stage data for a range of hydrological units: river stretches, lakes, floodplains and irrigated areas. This method illustrated how large amounts of MODIS images may be exploited to monitor flood dynamics with adequate spatial and temporal resolution and good accuracy, except during the flood rise due to cloud presence. Previous correlations between flow levels and flooded areas were refined to account for the hysteresis as the flood recedes and for the varying amplitude of the flood. Peak flooded areas varied between 10 300 km2 and 20 000 km2, resulting in evaporation losses ranging between 12 km3 and 21 km3. Direct precipitation assessed over flooded areas refined the wetland’s water balance and infiltration estimates. The knowledge gained on the timing, duration and extent of the flood across the wetland and in lakes, floodplains and irrigated plots may assist farmers in agricultural water management. Furthermore insights provided on the wetland’s flood dynamics may be used to develop and calibrate a hydraulic model of the flood in the Niger Inner Delta.

Impacts of land-use change on sacred forests at the landscape scale

Sacred forests often exist as isolated patches of natural forest even after conversion of the surrounding matrix to different forms of land-use. This study set out to: (1) evaluate land-cover changes and patch fragmentation in a landscape containing sacred and non-sacred forest patches over 15 years and (2) compare the effects at an individual patch level between sacred and non-sacred forests. Past changes in area and patch fragmentation of land cover classes and individual forest patches in the Gamo Highlands, Ethiopia, were assessed using maximum-likelihood classification of LANDSAT images. Large changes in land-cover occurred during 1995–2010, with 109.4% increase in area of farm and settlement and 36.6% decrease of forest area, with a decrease in number of forest patches by 16.1%, mean size by 26.8%, edge density by 29.1% and shape index by 13.3%. While all four individually studied non-sacred forests decreased in size over this period only four of the six individual sacred forests patches showed reduction in area. Forest patches with sacred status had greater protection by local communities than non-sacred forests in the Gamo Highlands. However, their small size and increasing edge density indicate high vulnerability, especially if an erosion of traditional cultural values reduces their protection.