Major Land Cover Research Articles

Land Cover and Land Use Change in the US Prairie Pothole Region Using the USDA Cropland Data Layer

The Prairie Pothole Region (PPR) is a biotically important region of grassland, wetland, and cropland that traverses the Canada-US border. Significant amounts of grasslands and wetlands within the PPR have been converted to croplands in recent years due to increasing demand for biofuels. We characterized land dynamics across the US portion of the PPR (US–PPR) using the USDA Crop Data Layer (CDL) for 2006–2018. We also conducted a comparative analysis between two epochs (1998–2007 & 2008–2017) of the CDL data time series in the North Dakotan portion of the US–PPR. The CDL revealed the western parts of the US–PPR have been dominated by grass/pasture, to the north it was spring wheat, to the east and southern half, soybeans dominated, and to the south it was corn (maize). Nonparametric trend analysis on the major crop and land cover types revealed statistically significant net decreases in the grass/pasture class between 2006 and 2018, which accounts for more than a quarter of grass/pasture area within the US–PPR. Other crops experiencing significant decreases included sunflower (-5%), winter wheat (-3%), spring wheat (-2%), and durum wheat (-1%). The combined coverage of corn and soybeans exhibited significant net increases in 23.5% of its cover; whereas, the individual significant net increases were 5% for corn and 11% for soybeans. Hotspots of increase in corn and soybeans were distributed across North and South Dakota. Other crop/land covers with huge significant increases include other hay/non-alfalfa (15%), and alfalfa (11%), which appear to be associated with the sharp increase in larger dairy operations, mostly in Minnesota. Wetland area increased 5% in the US–PPR, due to increased precipitation as well as inundation associated with Devils Lake in North Dakota. Hotspots of decreasing grass/pasture area were evident across the study area. Comparative trend analysis of two epochs (1998–2007 vs. 2008–2017) in North Dakota revealed that grass/pasture cover showed a negligible net trend (-0.3 %) between 1998 and 2007; whereas, there was a statistically significant decrease of more than 30% between 2008 and 2017. Combined coverage of corn and soybeans experienced statistically significant net increases in both epochs: 11% greater during 1998–2007 and 17% greater during 2008–2017. Recent sharp losses of grasslands and smaller wetlands combined the expansion of corn, soybeans, and alfalfa bode ill for wildlife habitat and require a re-examination of agricultural and energy policies that have encouraged these land transitions.

Global rangeland production systems and livelihoods at threat under climate change and variability

Rangelands are one of the Earth’s major ice-free land cover types. They provide food and support livelihoods for millions of people in addition to delivering important ecosystems services. However, rangelands are at threat from climate change, although the extent and magnitude of the potential impacts are poorly understood. Any declines in vegetation biomass and fluctuations in grazing availability would be of concern for food production and ecosystem integrity and functionality. In this study, we use a global rangeland model in combination with livestock and socio-economic datasets to identify where and to what extent rangeland systems may be at climatic risk. Overall, mean herbaceous biomass is projected to decrease across global rangelands between 2000 and 2050 under RCP 8.5 (−4.7%), while inter- (year-to-year) and intra- (month-to-month) annual variabilities are projected to increase (+21.3% and +8.2%, respectively). These averaged global estimates mask large spatial heterogeneities, with 74% of global rangeland area projected to experience a decline in mean biomass, 64% an increase in inter-annual variability and 54% an increase in intra-annual variability. Half of global rangeland areas are projected to experience simultaneously a decrease in mean biomass and an increase in inter-annual variability—vegetation trends both potentially harmful for livestock production. These regions include notably the Sahel, Australia, Mongolia, China, Uzbekistan and Turkmenistan and support 376 million people and 174 million ruminant Tropical Livestock Units. Additionally, the rangeland communities currently the most vulnerable (here, with the lowest livestock productivities and economic development levels and with the highest projected increases in human population densities) are projected to also experience the most damaging vegetation trends for livestock production. Although the capacity of rangeland systems to adapt is highly complex, analyses such as these generate some of the information required to inform options to facilitate pastoral system mitigation and adaptation strategies under climate change.

Spatiotemporal land cover dynamics and drivers for Dhidhessa River Basin (DRB), Ethiopia

Land cover change drives scale-dependent impacts on the environment. Effective environmental resources management, therefore, requires an understanding of the spatiotemporal land cover dynamics and their impacts at watershed scale. However, an integrated study that includes spatiotemporal land cover dynamics and the relative importance of drivers of the changes is missing for many basins in Ethiopia including the Dhidhessa River Basin (DRB). The objective of this study is to quantify historic and future land cover dynamics and to identify the relative importance of the drivers of the changes for the DRB. The post-classification comparison method and the cellular automata–Markov chain models were used for the analysis. The major land cover change driving factors were analyzed using the partial least square regression technique. The result revealed increasing waterbody, settlement, grassland, and agricultural lands at the expense of forest, and bush and shrubland in the past and future 30 years. Population density, slope, elevation, temperature, and rainfall were the drivers of land cover changes where population pressure is the primary driver. The observed land cover changes in the DRB could result in environmental degradation. Quantifying these impacts is, therefore, crucial to plan and implement informed land resources management strategies.

Introducing APiC for regionalised land cover mapping on the national scale using Sentinel-2A imagery

Overcoming the obstacle of frequent cloud coverage in optical remote sensing data is essential for monitoring dynamic land surface processes from space. APiC, a novel adaptable pixel-based compositing and classification approach, is especially designed to use high resolution spatio-temporal space-borne data.Here, pixel-based compositing is used separately for training data and prediction data. First, cloud-free pixels covered by reference data are used within adapted composite periods to compile a training dataset. The compiled training dataset contains samples of spectral reflectances for respective land cover classes at each composite period. For land cover prediction, pixel-based compositing is then applied region-wide. Multiple prediction models are used based on temporal subsets of the compiled training dataset to dynamically account for cloud coverage at pixel level. Thus we present a data-driven classification approach which is applicable in regions with different weather conditions, species composition and phenology.The capability of our method is demonstrated by mapping 19 land cover classes across Germany for the year 2016 based on Sentinel-2A data. Since climatic conditions and thus plant phenology change on a large scale, the classification was carried out separately in six landscape regions of different biogeographical characteristics. The study drew on extensive ground validation data provided by the federal states of Germany.For each landscape region, composite periods of different lengths have been established, which differ regionally in their temporal arrangement as well as in their total number, emphasising the advantage of a flexible regionalised classification procedure. Using a random forest classifier and evaluating outcomes with independent reference data, an overall accuracy of 88% was achieved, with particularly high classification accuracy of around 90% for the major land cover types. We found that class imbalances have significant influence on classification accuracy. Based on multiple temporal subsets of the compiled training dataset, over 10,000 random forest models were calculated and their performance varied considerably across and within landscape regions. The calculated importance of composite periods show that a high temporal resolution of the compiled training dataset is necessary to better capture the different phenology of land cover types.In this study we demonstrate that APiC, due to its data-driven nature, is a very flexible compositing and classification approach making efficient use of dense satellite time series in areas with frequent cloud coverage. Hence, regionalisation can be given greater focus in future broad-scale classifications in order to facilitate better integration of small-scale biophysical conditions and achieve even better results in detailed land cover mapping.

Land cover drives amphibian diversity across steep elevational gradients in an isolated neotropical mountain range: Implications for community conservation

Over 70% of land in the tropics is in some form of agricultural matrix which poses a threat to biodiversity. In Colombia, montane regions are dominated by varying intensities of agriculture and high levels of biodiversity and endemism. Globally, Colombia has the second largest number of amphibian species and is also the third largest coffee producer. Our study region, the Sierra Nevada de Santa Marta (SNSM), has high levels of amphibian endemism (38% and 10 threatened endemics) and is the fourth largest coffee growing region in Colombia. The SNSM rises from the sea to 5,775 m in just 42 km, with a direct overlap between coffee and amphibian habitat occurring across 600–1,800 m. We examined how land cover and elevation (from 800 to 3,700 m asl) influenced amphibian community structure, species richness, and abundance. We conducted surveys from September 2017 to July 2018 at 35 transects across five major land cover types: forest, ecotone, páramo, pasture, and shade coffee. In total, we recorded 19 species (366 individuals; 16 endemic species). Land cover was the main determinant of amphibian community structure, while the interaction between elevation and land cover was the main determinant of species richness and abundance. Forest and ecotone contained 73% of overall richness (14 species) with one species found exclusively in ecotone and three exclusively in forest. Pasture and coffee supported 42% (8 species) of species with only two species found exclusively at these land cover types. Shade coffee had low species richness and abundance and we detected just one endemic species in this land cover. The preservation of ecotone, transitional degraded habitat that occurs between two or more types of contiguous land cover types, represents an opportunity to safeguard microhabitats and microclimates. Conservation in the region should be collaborative and include private landowners, NGO’s, government agencies, and academics. Conservation actions should prioritize protecting extant natural habitat, restoring degraded habitats, increasing the heterogeneity of production systems, and improving landscape connectivity and watershed health. To achieve those actions, local communities will require economic incentives to maintain forest cover and reduce the contamination of streams through agricultural runoff.

Drivers of regional soil water storage memory and persistence

AbstractSoil water storage dynamics link landscape and climate processes with consequent impacts on hydrologic and biogeochemical cycles, and on ecosystem functions. We explore regional‐scale, spatiotemporal dynamics of root‐zone soil water storage using spatially explicit, long‐term simulations from a land data assimilation system over an area of ∼2 million km2 in the U.S. Midwest. We synthesize the dataset to examine spatial patterns of hydroclimatic forcing and soil and vegetation properties to identify the key drivers of soil water storage across the region. To identify differences in temporal hydrologic variability between sites with different soil and vegetation characteristics, we focus on daily data from 10 locations, representative of three major land cover types in the region. We use the concepts of (a) soil water memory to evaluate differences in landscape buffering of rainfall and (b) persistence to evaluate the threshold‐crossing properties of statistically defined “wet” and “dry” soil water conditions. Power spectral analyses of soil water storage reveal that regionally consistent patterns emerge in memory at multiple temporal scales. Threshold‐crossing analyses reveal corresponding similarity in persistence between sites. The analyses show that stochastic rainfall is the key driver of landscape hydrologic dynamics, with rainfall frequency as the primary determinant of persistence across the region, whereas differences in land cover and soil properties across the region have second‐order influence. The synthesis approach we present here is useful for ecological and agricultural risk assessments of climate change impacts, including increasing frequency of extreme events such as droughts and floods, but primarily points to the need for better understanding of future precipitation changes.

Mapping fire regime ecoregions in California

The fire regime is a central framing concept in wildfire science and ecology and describes how a range of wildfire characteristics vary geographically over time. Understanding and mapping fire regimes is important for guiding appropriate management and risk reduction strategies and for informing research on drivers of global change and altered fire patterns. Most efforts to spatially delineate fire regimes have been conducted by identifying natural groupings of fire parameters based on available historical fire data. This can result in classes with similar fire characteristics but wide differences in ecosystem types. We took a different approach and defined fire regime ecoregions for California to better align with ecosystem types, without using fire as part of the definition. We used an unsupervised classification algorithm to segregate the state into spatial clusters based on distinctive biophysical and anthropogenic attributes that drive fire regimes – and then used historical fire data to evaluate the ecoregions. The fire regime ecoregion map corresponded well with the major land cover types of the state and provided clear separation of historical patterns in fire frequency and size, with lower variability in fire severity. This methodology could be used for mapping fire regimes in other regions with limited historical fire data or forecasting future fire regimes based on expected changes in biophysical characteristics.

Patterns of invasion by crofton weed (Ageratina adenophora) in Kailash sacred landscape region of western Himalaya (India)

Invasion of alien species in high altitude ecosystems is a potent threat to the species diversity as well as it can cause severe environmental and economic issues. The invasion of alien plant species can be facilitated by many eco-climatic factors. The present study was conducted to assess patterns and trends of invasion by Ageratina adenophora in major land use and land cover types; in Gokerneshwergad watershed of Kailash Sacred Landscape (KSL) in western Himalaya. Extensive surveys were conducted to map the species in each season and habitat type. Sites with high biotic pressure and open forest canopy were the most suitable habitats for its growth. A negative correlation was found between distribution and altitude. The highest invasion was recorded in between 1700 – 1800m elevation gradient, between 20° and 30° slope positions and at North (33.33%), whereas, the lowest invasion was recorded between 700 – 800m in South-East directions (3.70%). Several other parameters such as distance from the disturbance site such as road, villages or settlements, drainage and soil texture were also found to be affecting the distribution pattern of this species. Interestingly results reveal that the alien plants also start competing among themselves after reaching their threshold level.

A Comparative Study of various Indices for extraction urban impervious surface of Landsat 8 OLI

Impervious surface is one of the major land cover types of urban and suburban environment. Conversion of rural landscapes and vegetation area to urban and suburban land use is directly related to the increase of the impervious surface area. The impervious surface expansion is straight-lined with decreasing green spaces in urban areas. Impervious surface is one of indicator for detecting urban heat islands. This study compares various indices for mapping impervious surfaces using Landsat 8 OLI imagery by optimizing the different spectral characteristics of Landsat 8 OLI imagery. The research objectives are (1) to apply various indices for impervious surface mapping and (2) identifies impervious surfaces in urban areas based on multiple indices and provide recommendations and find the best index for mapping impervious surface in urban areas. In addition to utilizing the index, land use supervised classification method, maximum likelihood classification used for extracting built-up, and non-built-up areas. Accuracy assessment of this research used field data collection as primary data for calculating kappa coefficient, producer accuracy, and user accuracy. The study can also be extended to find the land surface temperature and correlate the impervious surface extraction data with urban heat islands.

ANALYZING THE EFFECTS OF LAND COVER CHANGE ON SURFACE TEMPERATURE IN MOUNT MAKILING FOREST RESERVE (MMFR) AND ITS NEIGHBORING MUNICIPALITIES USING LANDSAT DATA

Abstract. Land Surface Temperature (LST) is said to be affected by frequent changes in the land cover. Over the years, the immediate environs of Mount Makiling Forest Reserve (MMFR) have experienced such kind of change due to rapid economic growth of the area that also led to the expansion of urban centers. The study utilized Landsat imageries to determine the possible effects of land cover change on surface temperature using the integration of remote sensing and GIS technologies. Initially, the multispectral bands were radiometrically corrected using Dark Object Subtraction (DOS) while the thermal bands were corrected using Land Surface Emissivity (LSE). After these corrections were applied, the images were classified using supervised image classification technique where seven land cover types have been identified. The classified images were then validated using 200 reference data and this revealed an overall accuracy of 87.5% and 86.0% for the May 2003 and July 2015 images, respectively. Results showed that changes in land cover resulted to a significant change in Land Surface Temperature (LST). The LST in 2003 (16.49°C – 40.44°C) was found higher than that of 2015 which was observed between 13.35°C and 33.83°C only. The reason behind this is the increase in green spaces from 2003 to 2015. Among the major land cover types, forest lands exhibited the lowest mean surface temperature for both years having 27.27°C in 2003 and 21.35°C in 2015 while built-up areas had the highest surface temperature having 32.60°C in 2003 and 26.00°C in 2015.

Mapping Urban Slum Settlements Using Very High-Resolution Imagery and Land Boundary Data

Accurate mapping of slums is crucial for urban planning and management. This article proposes a machine learning, hierarchical object-based method to map slum settlements using very high-resolution (VHR) imagery and land boundary data to support slum upgrading. The proposed method is tested in Kingston Metropolitan Area, Jamaica. First, the VHR imagery is classified into major land cover classes (i.e., the initial land cover map). Second, the VHR imagery and land boundary layer are used to obtain homogenous neighborhoods (HNs). Third, the initial land cover map is used to derive multiple context, spectral, and texture image features according to the local physical characteristics of slum settlements. Fourth, a machine-learning classifier, classification and regression trees, is used to classify HNs into slum and nonslum settlements using only the effective image features. Finally, reference data collected manually are used to assess the accuracy of the classification. In the training site, an overall accuracy of 0.935 is achieved. The effective image indicators for slum mapping include the building layout, building density, building roof characteristics, and distance from buildings to gullies. The classifier and those features selected from the training site are further used to map slums in two validating sites to assess the transferability of our approach. Overall accuracy of the two validating sites reached 0.928 and 0.929, respectively, suggesting that the features and classification model obtained from one site has the potential to be transferred to other areas in Jamaica and possibly other developing Caribbean countries with similar situation and data availability.

Analyzing the Effects of Temporal Resolution and Classification Confidence for Modeling Land Cover Change with Long Short-Term Memory Networks

Land cover change (LCC) is typically characterized by infrequent changes over space and time. Data-driven methods such as deep learning (DL) approaches have proven effective in many domains for predictive and classification tasks. When applied to geospatial data, sequential DL methods such as long short-term memory (LSTM) have yielded promising results in remote sensing and GIScience studies. However, the characteristics of geospatial datasets selected for use with these methods have demonstrated important implications on method performance. The number of data layers available, the rate of LCC, and inherent errors resulting from classification procedures are expected to influence model performance. Yet, it is unknown how these can affect compatibility with the LSTM method. As such, the main objective of this study is to explore the capacity of LSTM to forecast patterns that have emerged from LCC dynamics given varying temporal resolutions, persistent land cover classes, and auxiliary data layers pertaining to classification confidence. Stacked LSTM modeling approaches are applied to 17-year MODIS land cover datasets focused on the province of British Columbia, Canada. This geospatial data is reclassified to four major land cover (LC) classes during pre-processing procedures. The evaluation considers the dataset at variable temporal resolutions to demonstrate the significance of geospatial data characteristics on LSTM method performance in several scenarios. Results indicate that LSTM can be utilized for forecasting LCC patterns when there are few limitations on temporal intervals of the datasets provided. Likewise, this study demonstrates improved performance measures when there are classes that do not change. Furthermore, providing classification confidence data as ancillary input also demonstrated improved results when the number of timesteps or temporal resolution is limited. This study contributes to future applications of DL and LSTM methods for forecasting LCC.

Habitat and Nest Site Selection of the European Rollers in the Northern Anatolia

In this study, habitat preferences of the European Rollers (Coracias garrulus) in the Northern Anatolia was investigated, in 2014 breeding season. Total of 31 breeding pairs in 4 certain roller sites were located. The average distance between the breeding territories was 1.2 km in one of the roller sites with 17 pairs. Both solitary and colonial breeding pairs were recorded. Overall, 8, 2 and 2 nests (total of 12 nests) were located on sand quarries, concrete buildings and abandoned magpie nests on electrical poles, respectively. The habitat preferences were determined within a radius of 300 m and 1000 m around the nest site for the breeding and foraging habitats, respectively. We conducted the spatial analysis in ArcGIS 10.5 using Corine Land Cover 2012 database. Agricultural areas were the major land cover types where the European Rollers breed in our study. Use of unfavorable habitats and other nest sites including magpie nests necessitated a conservation program for the European Rollers in the study area.

Satellite-detected gain in built-up area as a leading economic indicator

Leading indicators of future economic activity include measures such as new housing starts, managers purchasing index, money supply, and bond yields. Such macroeconomic and financial indicators hold predictive power in signaling recessionary periods. However, many indicators are constrained by the fact that data are often published with some delay and are subject to constant revision (Bandholz and Funke , Huang et al , Orphanides ). In this research, we propose a leading indicator derived from satellite imagery, the expansion of anthropogenic bare ground. Satellite-detected gain in built-up area, a major land cover and land use (LCLU) outcome of anthropogenic bare ground gain (ABGG), provides an inexpensive, consistent, and near-real-time indicator of global and regional macroeconomic change. Our panel data analysis across four major regions of the world from 2001 to 2012 shows that the logarithm of total ABGG, mostly owing to its major LCLU outcome, the expansion of built-up land in either year t, t −1 or t −2, significantly correlated with the year t logarithm of gross domestic product (GDP, de-trended by Hodrick–Prescott filter). Global ABGG between 2001 and 2012 averaged 7875 km2 yr−1, with a peak gain of 11 875 (± 2014 km2 at the 95% confidence interval) in 2006, prior to the 2007–2008 global financial crisis. The curve of global ABGG or its major LCLU outcome of built-up area in year t − 1 accords well with that of the de-trended logarithm of the global GDP in year t. Given the 40 year archive of free satellite data, a growing satellite constellation, advances in machine learning, and scalable methods, this study suggests that analyses of ABGG as a whole or its LCLU outcomes can provide valuable information in near-real time for socioeconomic research, development planning, and economic forecasting.

Groundwater temperature anomalies in central Europe

As groundwater is competitively used for drinking, irrigation, industrial and geothermal applications, the focus on elevated groundwater temperature (GWT) affecting the sustainable use of this resource increases. Hence, in this study GWT anomalies and their heat sources are identified. The anthropogenic heat intensity (AHI), defined as the difference between GWT at the well location and the median of surrounding rural background GWTs, is evaluated in over 10 000 wells in ten European countries. Wells within the upper three percentiles of the AHI are investigated for each of the three major land cover classes (natural, agricultural and artificial). Extreme GWTs ranging between 25 °C and 47 °C are attributed to natural hot springs. In contrast, AHIs from 3 to 10 K for both natural and agricultural surfaces are due to anthropogenic sources such as landfills, wastewater treatment plants or mining. Two-thirds of all anomalies beneath artificial surfaces have an AHI > 6 K and are related to underground car parks, heated basements and district heating systems. In some wells, the GWT exceeds current threshold values for open geothermal systems. Consequently, a holistic management of groundwater, addressing a multitude of different heat sources, is required to balance the conflict between groundwater quality for drinking and groundwater as an energy source or storage media for geothermal systems.

Modelling the effects of permafrost loss on discharge from a wetland‐dominated, discontinuous permafrost basin

AbstractPermafrost degradation in the peat‐rich southern fringe of the discontinuous permafrost zone is catalysing substantial changes to land cover with expansion of permafrost‐free wetlands (bogs and fens) and shrinkage of forest‐dominated permafrost peat plateaux. Predicting discharge from headwater basins in this region depends upon understanding and numerically representing the interactions between storage and discharge within and between the major land cover types and how these interactions are changing. To better understand the implications of advanced permafrost thaw‐induced land cover change on wetland discharge, with all landscape features capable of contributing to drainage networks, the hydrological behaviour of a channel fen sub‐basin in the headwaters of Scotty Creek, Northwest Territories, Canada, dominated by peat plateau–bog complexes, was modelled using the Cold Regions Hydrological Modelling platform for the period of 2009 to 2015. The model construction was based on field water balance observations, and performance was deemed adequate when evaluated against measured water balance components. A sensitivity analysis was conducted to assess the impact of progressive permafrost loss on discharge from the sub‐basin, in which all units of the sub‐basin have the potential to contribute to the drainage network, by incrementally reducing the ratio of wetland to plateau in the modelled sub‐basin. Simulated reductions in permafrost extent decreased total annual discharge from the channel fen by 2.5% for every 10% decrease in permafrost area due to increased surface storage capacity, reduced run‐off efficiency, and increased landscape evapotranspiration. Runoff ratios for the fen hydrological response unit dropped from 0.54 to 0.48 after the simulated 50% permafrost area loss with a substantial reduction of 0.47 to 0.31 during the snowmelt season. The reduction in peat plateau area resulted in decreased seasonal variability in discharge due to changes in the flow path routing, with amplified low flows associated with small increases in subsurface discharge, and decreased peak discharge with large reductions in surface run‐off.

Operational Flood Mapping Using Multi-Temporal Sentinel-1 SAR Images: A Case Study from Bangladesh

Bangladesh is one of the most flood-affected countries in the world. In the last few decades, flood frequency, intensity, duration, and devastation have increased in Bangladesh. Identifying flood-damaged areas is highly essential for an effective flood response. This study aimed at developing an operational methodology for rapid flood inundation and potential flood damaged area mapping to support a quick and effective event response. Sentinel-1 images from March, April, June, and August 2017 were used to generate inundation extents of the corresponding months. The 2017 pre-flood land cover maps were prepared using Landsat-8 images to identify major land cover on the ground before flooding. The overall accuracy of flood inundation mapping was 96.44% and the accuracy of the land cover map was 87.51%. The total flood inundated area corresponded to 2.01%, 4.53%, and 7.01% for the months April, June, and August 2017, respectively. Based on the Landsat-8 derived land cover information, the study determined that cropland damaged by floods was 1.51% in April, 3.46% in June, 5.30% in August, located mostly in the Sylhet and Rangpur divisions. Finally, flood inundation maps were distributed to the broader user community to aid in hazard response. The data and methodology of the study can be replicated for every year to map flooding in Bangladesh.

Determination of Vegetation Thresholds for Assessing Land Use and Land Use Changes in Cambodia using the Google Earth Engine Cloud-Computing Platform

As more data and technologies become available, it is important that a simple method is developed for the assessment of land use changes because of the global need to understand the potential climate mitigation that could result from a reduction in deforestation and forest degradation in the tropics. Here, we determined the threshold values of vegetation types to classify land use categories in Cambodia through the analysis of phenological behaviors and the development of a robust phenology-based threshold classification (PBTC) method for the mapping and long-term monitoring of land cover changes. We accessed 2199 Landsat collections using Google Earth Engine (GEE) and applied the Enhanced Vegetation Index (EVI) and harmonic regression methods to identify phenological behaviors of land cover categories during the leaf-shedding phenology (LSP) and leaf-flushing phenology (LFS) seasons. We then generated 722 mean phenology EVI profiles for 12 major land cover categories and determined the threshold values for selected land cover categories in the mid-LSP season. The PBTC pixel-based classified map was validated using very high-resolution (VHR) imagery. We obtained a cumulative overall accuracy of more than 88% and a cumulative overall accuracy of the referenced forest cover of almost 85%. These high accuracy values suggest that the very first PBTC map can be useful for estimating the activity data, which are critically needed to assess land use changes and related carbon emissions under the Reducing Emissions from Deforestation and forest Degradation (REDD+) scheme. We found that GEE cloud-computing is an appropriate tool to use to access remote sensing big data at scale and at no cost.

Learning from data: A post classification method for annual land cover analysis in urban areas

Annual analyses of land cover dynamics in urban areas provide a thorough understanding of the urbanization effects on environment and valuable information for the improvement of urban growth modeling. However, most current studies focus on major land cover changes, such as urbanization and vegetation loss. The most feasible way to evaluate the complex interactions among different land cover types is the post-classification change detection, but the temporal inconsistency in the time series of land cover maps impedes the high-frequency and long-term analyses. This study proposed a spatio-temporal land cover filter (STLCF) to remove the illogical land cover change events in the time series of land cover maps, and analyzed the annual land cover dynamics in urban areas. The knowledge of illogical land cover change events was ‘learned’ from the land cover maps through the spatio-temporal transition probability matrix, instead of experts’ knowledge. The illogical change was modified with the land cover of the maximum probability calculated from the naïve Bayesian equation. The STLCF was tested in Wuhan, a typical densely urbanized Chinese city. The annual land cover maps from 2000 to 2013 were derived from multi-date Landsat images using the Decision Tree (DT) classifier. Results showed that the STLCF improved the mean overall accuracy of annual change detection by about 6%. Additionally, the amount of land cover trajectories with unrealistically frequent changes was significantly decreased. During the study period, 7.86% of the pixels experienced one land cover change, and about 0.57% of the pixels experienced land cover changes more than once. The annual analyses demonstrated the non-linear increasing trend in urbanization as well as the corresponding trend in vegetation loss in the study area. We also found the conversion from built-up areas to vegetation near rivers and lakes and in the reserves and rural areas, mainly caused by the restoration of built-up areas to the park or green belt/wedges along rivers and new roads in the metropolitan areas, and to the cropland and woods in the rural areas. Results of this study showed the importance of the spatio-temporal consistency check with knowledge derived from land cover maps of the study area, which facilitates the annual analyses of major and subtle land cover dynamics in urban areas.

Mapping of burned area and assessment of burn severity of the 2017 wild fires in Kresna gorge

At the end of the summer of 2017, the weather in Bulgaria was very dry and hot and this contributed for the occurrence of a series of wild fires. The largest and most prolonged was the wild fire that raged in Kresna gorge. The fire broke out on August 24 and lasted for nearly a week. The aim of the paper is to make an assessment of the damage caused by the August fire and by a smaller wild fire in the same region, which broke out on July 13, 2017. The assessment is based on images from the satellites Sentinel-2A and -2B. For the identification and mapping of burned areas the spectral indices NBR (Normalized Burn Ratio) and dNBR (difference NBR) were used. Map of burn severity within the fire perimeter was compiled by reclassifying the dNBR values into three classes. The total burned area was 2199.1 ha; of this, the August fire covered 96%. The greatest part of the fire scar, over 50% of the area, is characterized with medium burn severity (dNBR between 0.27 and 0.66). Both, the lowest (dNBR between 0.1 and 0.27) and highest (dNBR>0.66) burn severity levels covered about ¼ of the total burned area. A map of the major land cover types in the study area was prepared in order to assess the most severely affected land cover types. The map is derived from supervised SVM (Support Vector Machine) classification of a Sentinel-2B image acquired before the fires. It was found that the coniferous forest class is dominated by the highest burn severity level (43.7% of the area). In the broadleaf forest class the areas with the lowest level of burn severity prevail (53.5%). The main reason is that coniferous forests in the study area represent artificial plantations created at altitude lower than the usual altitude of coniferous species in the region. Therefore they are much more susceptible to fire than broadleaf species which are native to the oak belt.