Multi-temporal Satellite Data Research Articles

Land Cover Mapping with Convolutional Neural Networks Using Sentinel-2 Images: Case Study of Rome

Land cover monitoring is crucial to understand land transformations at a global, regional and local level, and the development of innovative methodologies is necessary in order to define appropriate policies and land management practices. Deep learning techniques have recently been demonstrated as a useful method for land cover mapping through the classification of remote sensing imagery. This research aims to test and compare the predictive models created using the convolutional neural networks (CNNs) VGG16, DenseNet121 and ResNet50 on multitemporal and single-date Sentinel-2 satellite data. The most promising model was the VGG16 both with single-date and multi-temporal images, which reach an overall accuracy of 71% and which was used to produce an automatically generated EAGLE-compliant land cover map of Rome for 2019. The methodology is part of the land mapping activities of ISPRA and exploits its main products as input and support data. In this sense, it is a first attempt to develop a high-update-frequency land cover classification tool for dynamic areas to be integrated in the framework of the ISPRA monitoring activities for the Italian territory.

The effect of post-wildfire management practices on vegetation recovery: Insights from the Sapadere fire, Antalya, Türkiye

Post-wildfire management actions mainly targeting the removal of salvage logs and burned trees is a common but controversial practice. Although it aims to regain some of the natural and economic value of a forest, it also requires disturbing burned areas, which may have some negative consequences affecting, for instance, the carbon cycle, soil erosion, and vegetation cover. Observations from different geographic settings contribute to this scientific debate, and yet, the spatiotemporal evolution of the post-fire road network developed as part of fire management practices and its influence on vegetation recovery has been rarely examined. Specifically, we still lack observations from Türkiye, though wildfires are a common event. This research examined the evolution of the vegetation cover in relation to post-fire road constructions and the resultant debris materials in areas affected by the 2017 Sapadere fire in Antalya, Türkiye. We used multi-sensor, multi-temporal optical satellite data and monitored the variation in both vegetation cover and road network from the pre-to post-fire periods between 2014 and 2021. Our results showed that fire management practices almost doubled the road network in the post-fire period, from 487 km to 900 km. Overall, 7% of the burned area was affected by these practices. As a result, vegetation cover in those areas shows only ∼50% recovery, whereas undisturbed areas exhibit ∼100% recovery 5 years after the event. Notably, such spatiotemporal analysis carried out for different burned areas would provide a better insight into the most suitable post-fire management practices. Our findings, in particular, show that the current practices need to be revisited as they cause a delay in vegetation recovery.

Spatio-temporal variations of a meandering channel morphodynamics: A study along the Kaljani River in Bhuchungmari-Moamari Section of North Bengal Dooars, West Bengal (India)

Rivers being the most dynamic and natural engines receive the maximum outcome of process-form activities over space and time. Displacement and shifting of course of the alluvial channels over the lower gradients result into severe bed-bank scouring activities temporally. River Kaljani within the neo tectonically active zone of North Bengal plains is an example of temporal variations of high sinuosity and braiding over the softer lithology. The present work is focusing on a temporal study from 1970 to 2019. The SOI Topographical map of 1:50,000 scales has been the base layer along with temporal satellite imageries (1990, 2000, 2010 and 2019). The schematization has followed digitization, rectification and overlaying of conformal layers extracted from the USGS website under UTM Projection system and WGS of 1984 datum for north 45 zones under GIS platform. Multi-temporal satellite data identified serious course changes of the selected reach of lower segment of Kaljani River. The rectified multi-spectral bands are used to make the NDWI for identification of the water content on earth surface followed by selective morphometric analysis e.g. sinuosity index, braided index and roundness of curvature to access the spatio-temporal changes accomplishing some derivative mathematical equations. Descriptive statistics viz. mean, standard deviation (SD), coefficient of variation (CV), skewness and kurtosis were used and explained maximum course oscillations. The authors tried to examine the fact for six temporal frames i.e. 1970, 1990, 2000, 2006, 2010 and 2019 and received high sinuosity of all categories. Conclusively the work claims to set a documentary approach to examine the morphodynamics of an alluvial channel of Sub-Himalayan foothills.

Monitoring and Modelling Morphological Changes in Rivers Using RS and GIS Techniques

River geomorphological investigation issues have received little attention in most countries of the world. Such processes become a pressing necessity due to climate change and anticipated events of extraordinary surges and dry seasons, which may debilitate the security of adjacent and downstream cities, particularly in locales that are exceedingly delicate and influenced by climatic changes. Al-Abbasia reach is a river that runs through the middle of the Euphrates River and is known for its numerous bends and meanders. The study of hydraulic structures such as barrages can provide important information about their influences on morphological processes in river reaches near the barrage upstream and downstream. Hydraulic analysis is made of the river behavior in u/s and d/s of hydraulic structures like barrages as a result of sediment deposition and erosion in u/s and d/s. A study, i.e., research on the impacts of the Abbassia barrage on the river system, has been conducted to address this issue using multi-temporal Landsat satellite data from 1976 to 2022 provided by the USGS. The study reach is located 5 kilometres upstream and 5 kilometres downstream of the Abbassia reach. Following the construction of the barrage, which had an impact on the sedimentation and geometry of the river, morphological variations took place in this part of the Al Abbassia reach. In this study, morphological changes throughout 49 years between 1976 and 2022 were investigated utilising remote sensing (RS) and geographic information system (GIS) approaches. Additionally, four image groups from three separate decades were used to perform change detection (1990–2000, 2000–2010, and 2010–2022). In this study, a monitoring system using Landsat-3 MSS: 1985, Landsat-5 TM: 1990, 1995, 2000, 2005, and Landsat-8 OLI: 2010, 2011, 2015, 2021, 2022 were employed to map river planform changes. The long-term comparison of this series of satellite images and historical maps for the period 1976–2022 indicates a continuation of change in the reach study with a rate of approximately 56, 33, 97, and 55% for upstream and 19%, 26%, 3%, and 45% for downstream for the width, area, deposition, and erosion, respectively. Furthermore, it is observed that there is a shift in river course within 200 m downstream of the barrage for the period of 1985–1990. The findings of this study, which monitor river morphological change at finer temporal and spatial resolutions, are crucial for promoting sustainable river management. They also aid in the investigation of river behaviour, which is necessary for providing the best management possible and overcoming the difficulties posed by this important research issue. Doi: 10.28991/CEJ-2023-09-03-03 Full Text: PDF

Mountain Tree Species Mapping Using Sentinel-2, PlanetScope, and Airborne HySpex Hyperspectral Imagery

Europe’s mountain forests, which are naturally valuable areas due to their high biodiversity and well-preserved natural characteristics, are experiencing major alterations, so an important component of monitoring is obtaining up-to-date information concerning species composition, extent, and location. An important aspect of mapping tree stands is the selection of remote sensing data that vary in temporal, spectral, and spatial resolution, as well as in open and commercial access. For the Tatra Mountains area, which is a unique alpine ecosystem in central Europe, we classified 13 woody species by iterative machine learning methods using random forest (RF) and support vector machine (SVM) algorithms of more than 1000 polygons collected in the field. For this task, we used free Sentinel-2 multitemporal satellite data (10 m pixel size, 12 spectral bands, and 21 acquisition dates), commercial PlanetScope data (3 m pixel size, 8 spectral bands, and 3 acquisitions dates), and airborne HySpex hyperspectral data (2 m pixel size, 430 spectral bands, and a single acquisition) with fusion of the data of topographic derivatives based on Shuttle Radar Topography Mission (SRTM) and airborne laser scanning (ALS) data. The iterative classification method achieved the highest F1-score with HySpex (0.95 RF; 0.92 SVM) imagery, but the multitemporal Sentinel-2 data cube, which consisted of 21 scenes, offered comparable results (0.93 RF; 0.89 SVM). The three images of the high-resolution PlanetScope produced slightly less accurate results (0.89 RF; 0.87 SVM).

Wetlands Mapping with Deep ResU-Net CNN and Open-Access Multisensor and Multitemporal Satellite Data in Alberta’s Parkland and Grassland Region

Wetlands are a valuable ecosystem that provides various services to flora and fauna. This study developed and compared deep and shallow learning models for wetland classification across the climatically dynamic landscape of Alberta’s Parkland and Grassland Natural Region. This approach to wetland mapping entailed exploring multi-temporal (combination of spring/summer and fall months over 4 years—2017 to 202) and multisensory (Sentinel 1 and 2 and Advanced Land Observing Satellite, ALOS) data as input in the predictive models. This input image consisted of S1 dual-polarization vertical-horizontal bands, S2 near-infrared and shortwave infrared bands, and ALOS-derived topographic wetness index. The study explored the ResU-Net deep learning (DL) model and two shallow learning models, namely random forest (RF) and support vector machine (SVM). We observed a significant increase in the average F1-score of the ResNet model prediction (0.82) compared to SVM and RF prediction of 0.69 and 0.69, respectively. The SVM and RF models showed a significant occurrence of mixed pixels, particularly marshes and swamps confused for upland classes (such as agricultural land). Overall, it was evident that the ResNet CNN predictions performed better than the SVM and RF models. The outcome of this study demonstrates the potential of the ResNet CNN model and exploiting open-access satellite imagery to generate credible products across large landscapes.

Optimal Sample Size and Composition for Crop Classification with Sen2-Agri’s Random Forest Classifier

Sen2-Agri is a software system that was developed to facilitate the use of multi-temporal satellite data for crop classification with a random forest (RF) classifier in an operational setting. It automatically ingests and processes Sentinel-2 and LandSat 8 images. Our goal was to provide practitioners with recommendations for the best sample size and composition. The study area was located in the Yaqui Valley in Mexico. Using polygons of more than 6000 labeled crop fields, we prepared data sets for training, in which the nine crops had an equal or proportional representation, called Equal or Ratio, respectively. Increasing the size of the training set improved the overall accuracy (OA). Gains became marginal once the total number of fields approximated 500 or 40 to 45 fields per crop type. Equal achieved slightly higher OAs than Ratio for a given number of fields. However, recall and F-scores of the individual crops tended to be higher for Ratio than for Equal. The high number of wheat fields in the Ratio scenarios, ranging from 275 to 2128, produced a more accurate classification of wheat than the maximal 80 fields of Equal. This resulted in a higher recall for wheat in the Ratio than in the Equal scenarios, which in turn limited the errors of commission of the non-wheat crops. Thus, a proportional representation of the crops in the training data is preferable and yields better accuracies, even for the minority crops.

A Hybrid Algorithm for Urban LULC Change Detection for Building Smart-city by Using WorldView Images

Technological advancement in smart cities can have adverse effects on the environment. Timely monitoring of smart cities to preserve environmental sustainability is a thrust area of research. It can be done by using change detection with multi-temporal satellite data. The success of these methods solely depends on the calibre of the backend image segmentation and Land-use Land-cover classification technique. The limitation of using cutting-edge classification algorithms is the availability of a proper dataset and identification of the edge structure of different LULC classes. In contrast, a segmentation algorithm cannot detect LULC classes automatically. In this research, we eliminated these shortcomings by considering a hybrid approach. We proposed a multi-class Support Vector Machine (SVM) and ISODATA-embedded large-scale change detection method. This method can automatically segment, detect, and perform LULC change analysis. We have considered the multi-sensor dataset of Barasat, West Bengal, India, obtained from the WorldView satellite sensor for the experimental study. The proposed method is validated concerning three different cutting-edge methods.

ASSESSMENT OF WETLAND DEGRADATION CONDITION AND TREND BASED ON LAND USE/COVER IN WAMI-RUVU RIVER BASIN USING MULTI-TEMPORAL LANDSAT IMAGES

The understanding of wetlands’ distribution and their level of susceptibility is important to enhance management and conservation efforts. The study aimed to map wetlands and assess their distribution pattern and their degradation based on the LULC by the years 2000, 2010 and 2020. Wetland types i.e., marshes, ponds, lakes, fens, rivers, floodplains, swamps, and open water bodies with other land use such as agriculture land, built-up and natural vegetation land were classified and mapped using random forest algorithm in Arc Map. Multi-temporal Landsat satellite data including thematic Mapper Landsat (+ETM) and Operational Land Imager (OLI) images were used for LULC mapping. The random forest classifier technique is used to analyse the satellite imagery for detecting the LULC changes during the whole study period. To assess the wetland area degradation three aspects were used in this study for the analysis; wetland area change rate, annual land cover change and annual land cover change rate. The study found about -34.41Km2/y of wetlands potentially are degraded annually due to human and natural stressors in 20 years of our study however the basin experienced a higher rate of wetland degradation in the second epoch of 2010-2020 where the basin experience -42.8Km2/y degradation which results to the total loss of -687.75 km2 total coverage of wetland area in the basin. This information could be used to improve wetland planning and management by wetland managers and other stakeholders.

The Application of Multi Temporal Satellite Data For Land Cover Mapping of Bawean Island, East Java

Land cover dynamics in a small island can be determined using Geographic Information System (GIS) approach based on multitemporal image analysis. This study aimed to classify major land cover types and to map land cover changes of Bawean Island. Two sets of 10 meter resolution satellite data ALOS AVNIR (2010) and Sentinel-2A (2020) were used in this study. Satellite image analysis was carried out through several stages namely image pre-processing including radiometric and geometric correction, supervised image classification and accuracy test. Image classification results from 2010 to 2020 showed a significant change in land cover on Bawean Island. The forest vegetation land cover declined significantly from 13,470.5 Ha in 2010 to 8,543.4 Ha in 2020. Most of the area have been converted into paddy fields and built-up areas. The accuracy test and validation were determined by comparing the 2020 Sentinel image classification results with field observation conducted in 2021. The analysis showed good results with 82.52% overall accuracy and 79.66 Kappa coefficient. Further investigation found that changes in land cover on Bawean Island occured due to the agriculture and infrastucture development.
 Keywords : Geographic Information System (GIS); land cover; satellite images; small island
 
 Copyright (c) 2022 Geosfera Indonesia and Department of Geography Education, University of Jember
 This work is licensed under a Creative Commons Attribution-Share A like 4.0 International License

Glacial Lake Outburst Flood Hazard and Risk Assessment of Gangabal Lake in the Upper Jhelum Basin of Kashmir Himalaya Using Geospatial Technology and Hydrodynamic Modeling

Climate warming-induced glacier recession has resulted in the development and rapid expansion of glacial lakes in the Himalayan region. The increased melting has enhanced the susceptibility for Glacial Lake Outburst Floods (GLOFs) in the region. The catastrophic failure of potentially dangerous glacial lakes could be detrimental to human life and infrastructure in the adjacent low-lying areas. This study attempts to assess the GLOF hazard of Gangabal lake, located in the Upper Jhelum basin of Kashmir Himalaya, using the combined approaches of remote sensing, GIS, and dam break modeling. The parameters, such as area change, ice thickness, mass balance, and surface velocity of the Harmukh glacier, which feeds Gangabal lake, were also assessed using multitemporal satellite data, GlabTop-2, and the Cosi–Corr model. In the worst-case scenario, 100% volume (73 × 106 m3) of water was considered to be released from the lake with a breach formation time (bf) of 40 min, breach width (bw) of 60 m, and producing peak discharge of 16,601.03 m3/s. Our results reveal that the lake area has increased from 1.42 km2 in 1972 to 1.46 km2 in 1981, 1.58 km2 in 1992, 1.61 km2 in 2001, 1.64 km2 in 2010, and 1.66 km2 in 2020. The lake area experienced 17 ± 2% growth from 1972 to 2020 at an annual rate of 0.005 km2. The feeding glacier (Harmukh) contrarily indicated a significant area loss of 0.7 ± 0.03 km2 from 1990 (3.36 km2) to 2020 (2.9 km2). The glacier has a maximum, minimum, and average depth of 85, 7.3, and 23.46 m, respectively. In contrast, the average velocity was estimated to be 3.2 m/yr with a maximum of 7 m/yr. The results obtained from DEM differencing show an average ice thickness loss of 11.04 ± 4.8 m for Harmukh glacier at the rate of 0.92 ± 0.40 m/yr between 2000 and 2012. Assessment of GLOF propagation in the worst-case scenario (scenario-1) revealed that the maximum flood depth varies between 3.87 and 68 m, the maximum flow velocity between 4 and 75 m/s, and the maximum water surface elevation varies between 1548 and 3536 m. The resultant flood wave in the worst-case scenario will reach the nearest location (Naranaag temple) within 90 min after breach initiation with a maximum discharge of 12,896.52 m3 s−1 and maximum flood depth and velocity of 10.54 m and 10.05 m/s, respectively. After evaluation of GLOF impacts on surrounding areas, the area under each inundated landuse class was estimated through the LULC map generated for both scenarios 1 and 2. In scenario 1, the total potentially inundated area was estimated as 5.3 km2, which is somewhat larger than 3.46 km2 in scenario 2. We suggest a location-specific comprehensive investigation of Gangbal lake and Harmukh glacier by applying the advanced hazard and risk assessment models/methods for better predicting a probable future GLOF event.

Monitoring of land use/land cover changes and its implications in the peri-urban areas using multi-temporal landsat satellite data: a case study of Guwahati city, Assam, India

Rapid population growth and urbanization in and around Guwahati city, the state capital of Assam have resulted in tremendous pressure on its limited land and civic infrastructures. Due to severe congestion and shortage of space, people are exploring peri-urban plain and hilly areas around the city and gradually expanding towards adjoining districts for settlement. Earlier studies focused mainly on Guwahati city only and neighboring districts, small and medium towns around the city, and adjacent Meghalaya state have received scant attention in urban policy research. This study evaluates land use land cover changes (LULC) from 1991 to 2021 within the GMDA area and encircling a 50 km buffer from the GMDA boundary. Landsat time series data were used to analyze land use and land cover changes. The result indicates the expansion in built-up areas with a net change of 2.27% from 1991 to 2021 within the 50 km buffer study area. The result indicates the loss of forest class and cultivated land with a net change percentage of 2.12 and 0.84% within the period of 30 years. Apart from the above, the study also reveals an increase in the built-up of 33.25% within the GMDA core area from 1991 to 2021. The impact of these LULC changes shows the expansion of the built-up area beyond the GMDA boundary in the eastern, western, and northwest direction and needs to be regulated for ensuring the planned growth of its outskirts and the surrounding region. Hence through adequate decentralized regional planning, the fast-expanding Guwahati city and its surrounding region can be managed.

Assessment of mangrove cover dynamics and its health status in the Gulf of Khambhat, Western India, using high-resolution multi-temporal satellite data and Google Earth Engine.

Anthropogenic activity is a major driving factor of greenhouse gas emission, leading to climate change worldwide. So, the best natural approach to lowering the carbon from the atmosphere is mangroves which have more potential to sequestrate carbon. But mangroves are under threat due to land use land cover change. This research has been carried out on the mangroves of Gulf of Khambhat, Gujarat, India, where anthropic activity is affecting the mangrove forest cover with spatiotemporal heterogeneity. In the present study, multi-temporal high-resolution satellite data AVNIR-2 (Advanced Visible and Near Infrared Radiometer type-2) and LISS-4 (Linear Imaging Self-Scanning Sensors-4) were used for the demarcation of various land use/land cover class (LULC), and change analysis and assessment of mangroves health for the years 2009, 2014, and 2019. The impact of saltpan/aquaculture on mangroves growth and its health status has been calculated by various MODIS (Moderate Resolution Imaging Spectroradiometer) satellite data products such as gross primary productivity (GPP), enhanced vegetation index (EVI), and leaf area index (LAI) in Google Earth Engine (GEE), and field-based method was also considered. This study suggests that there is a marginal increase (17.11 km2) in mangrove cover during the assessment period 2009-2019; on other side, 65.42 km2 was degraded also. However, increase in saltpan/aquaculture is imposing an adverse effect on mangroves' basal area, plant density, and productivity. Change analysis also suggests a reduction in healthy mangrove area (from 25.20 to 2.84 km2), which will have an impact on ecosystem services.

Estimating Groundnut Yield in Smallholder Agriculture Systems Using PlanetScope Data

Crop yield is related to household food security and community resilience, especially in smallholder agricultural systems. As such, it is crucial to accurately estimate within-season yield in order to provide critical information for farm management and decision making. Therefore, the primary objective of this paper is to assess the most appropriate method, indices, and growth stage for predicting the groundnut yield in smallholder agricultural systems in northern Malawi. We have estimated the yield of groundnut in two smallholder farms using the observed yield and vegetation indices (VIs), which were derived from multitemporal PlanetScope satellite data. Simple linear, multiple linear (MLR), and random forest (RF) regressions were applied for the prediction. The leave-one-out cross-validation method was used to validate the models. The results showed that (i) of the modelling approaches, the RF model using the five most important variables (RF5) was the best approach for predicting the groundnut yield, with a coefficient of determination (R2) of 0.96 and a root mean square error (RMSE) of 0.29 kg/ha, followed by the MLR model (R2 = 0.84, RMSE = 0.84 kg/ha); in addition, (ii) the best within-season stage to accurately predict groundnut yield is during the R5/beginning seed stage. The RF5 model was used to estimate the yield for four different farms. The estimated yields were compared with the total reported yields from the farms. The results revealed that the RF5 model generally accurately estimated the groundnut yields, with the margins of error ranging between 0.85% and 11%. The errors are within the post-harvest loss margins in Malawi. The results indicate that the observed yield and VIs, which were derived from open-source remote sensing data, can be applied to estimate yield in order to facilitate farming and food security planning.

Study of Fluctuations in Surface Area of Lake Haramaya using NDWI and MNDWI Methods

Lakes are of great value to human beings and important for various reasons like regulating the flow of river water, to maintain the eco-system and storage of water during the dry seasons. Lake Haramaya, which is situated at 14 km Northwest of Harer town (UNESCO Site) is one of the famous and beautiful lakes of Ethiopia. It acts as a source of life for human beings and animals. The over exploitation of the lake haramaya for water supply and agricultural purposes has led to its extinction in the last two decades. This study attempts to identify the fluctuations in surface area of Lake Haramaya between 1995 to 2020 by using multi-temporal satellite data. The Landsat 5TM images of 1995, 2000 and 2010, Landsat 7ETM+ image of 2005 and Landsat 8 OLI TIRS images of 2015 and 2020 are analyzed using the Normalized Difference Water Index (NDWI) and Modified Normalized Difference Water Index (MNDWI) methods. These methods are used to quantify the changes in surface area and compared to each other for identifying the suitable method for detecting water bodies. The present study shows that, between 1995 to 2010 the lake lost up to 2.3238 sq.km and almost dried. But, during 2011 to 2020 the lake surface area increased by 2.6946 sq.km. The study states that the lake surface area is fluctuating and MNDWI method is highly reliable in extracting water bodies.

LSTM-Based Prediction of Mediterranean Vegetation Dynamics Using NDVI Time-Series Data

Vegetation index time-series analysis of multitemporal satellite data is widely used to study vegetation dynamics in the present climate change era. This paper proposes a systematic methodology to predict the Normalized Difference Vegetation Index (NDVI) using time-series data extracted from the Moderate Resolution Imaging Spectroradiometer (MODIS). The key idea is to obtain accurate NDVI predictions by combining the merits of two effective computational intelligence techniques; namely, fuzzy clustering and long short-term memory (LSTM) neural networks under the framework of dynamic time warping (DTW) similarity measure. The study area is the Lesvos Island, located in the Aegean Sea, Greece, which is an insular environment in the Mediterranean coastal region. The algorithmic steps and the main contributions of the current work are described as follows. (1) A data reduction mechanism was applied to obtain a set of representative time series. (2) Since DTW is a similarity measure and not a distance, a multidimensional scaling approach was applied to transform the representative time series into points in a low-dimensional space, thus enabling the use of the Euclidean distance. (3) An efficient optimal fuzzy clustering scheme was implemented to obtain the optimal number of clusters that better described the underline distribution of the low-dimensional points. (4) The center of each cluster was mapped into time series, which were the mean of all representative time series that corresponded to the points belonging to that cluster. (5) Finally, the time series obtained in the last step were further processed in terms of LSTM neural networks. In particular, development and evaluation of the LSTM models was carried out considering a one-year period, i.e., 12 monthly time steps. The results indicate that the method identified unique time-series patterns of NDVI among different CORINE land-use/land-cover (LULC) types. The LSTM networks predicted the NDVI with root mean squared error (RMSE) ranging from 0.017 to 0.079. For the validation year of 2020, the difference between forecasted and actual NDVI was less than 0.1 in most of the study area. This study indicates that the synergy of the optimal fuzzy clustering based on DTW similarity of NDVI time-series data and the use of LSTM networks with clustered data can provide useful results for monitoring vegetation dynamics in fragmented Mediterranean ecosystems.

Mapping of crop types in 1989, 1999, 2009 and 2019 to assess major land cover trends of the Udon Thani Province, Thailand

Crop production and productivity monitoring play a crucial role for food security and livelihoods, international trade and sustainable agricultural practices. Earth Observation (EO) data provides high spectral, spatial and temporal data for various agricultural applications. However, mapping and monitoring small crop fields and complex landscapes are still challenging, in particular when attempting to trace the historical evolution of land use changes. To address this issue, a study was set up in the Udon Thani Province of Thailand, with small agricultural parcels and highly fragmented landscapes, covering an area of approximately 11,000 km2. Three decades of crop type dynamics were monitored and assessed using different combinations of multi-temporal Sentinel-1, Sentinel-2 and Landsat data and the random forest (RF) classifier. The combined multi-temporal EO datasets proved the most efficient for mapping crop types. Classification results achieved overall accuracy (OA) of 87.9%, 88.1%, 84.8% and 92.6% for the four base-years 1989, 1999, 2009 and 2019, respectively. Thanks to the availability of high-quality reference labels, the crop type map of 2019 showed the highest overall and class-specific accuracies. The 2019 classification model separated many crop classes well, especially sugarcane, cassava, rice and para rubber. On the contrary, for 1989, 1999 and 2009 drops in accuracy had to be accepted, as direct field reference observations were unavailable and reference information had to be sourced through photo-interpretation or trimming approaches. Overall, however, the RF method together with multi-temporal EO satellite data from multiple platforms showed high potential and excellent efficiency in crop type classification in complex landscapes. The most dominant classes of crop types for the four base-years were rice, sugarcane, and cassava, respectively. Land cover changes indicated that transitions of 1,529 km2 (14%) occurred between 1989 and 1999, mainly as increase in sugarcane and rice areas. From 1999 to 2009, significant land changes were observed covering 2,340 km2 (21%), primarily as increased cassava and para rubber cultivation. During the most recent period 2009 to 2019, an additional 3,414 km2 (31%) were transformed, mainly through the expansion of para rubber and sugarcane plantations. The main drivers for the observed land use changes in the three decades were commodity prices and agricultural policies. The cost-efficiently derived results provide valuable information to inform land use management decisions of policymakers and other stakeholders, including the consideration of environmental aspects.

Developing a spatial information system of biomass potential from crop residues over India: A decision support for planning and establishment of biofuel/biomass power plant

The study demonstrates effective utilization of geospatial technology to amalgamate crop statistics, satellite derived crop maps and productivity proxies to generate spatial maps of biomass potential from crop residues over a diverse country like India. District level crop production statistics were used to estimate gross biomass potential (GBP) of the selected crops (rice, wheat, sugarcane and cotton). Multi-temporal satellite data were used to generate crop maps. District level GBP was disaggregated at 1 km grid using satellite derived crop fraction and seasonal Gross Primary Production. The GBP maps were further converted to surplus biomass potential (SBP) followed by bioenergy potential using district specific surplus factors (ground survey) and crop specific calorific values respectively. These maps could capture the spatial variations of the available biomass resources and prospective zones to prioritize its utilization. The study found the presence of rice-wheat based SBP over Punjab and Haryana (>0.30 Ktonnes km−2), eastern Uttar Pradesh (0.15–0.30 Ktonnes km−2); sugarcane-wheat based SBP over western Uttar Pradesh (>0.25 Ktonnes km−2); cotton-wheat based SBP over the southern Haryana and western Punjab (>0.20 Ktonnes km−2); and sugarcane-cotton based SBP over central India and Gujarat (>0.20 Ktonnes km−2). Geospatial maps of crop residue biomass along with other thematic layers were hosted in a geoportal (BHUVAN-JAIVOORJA) with an online query module to assess the amount and types of biomass, land resources, logistics support based on user defined point of interest and fetch/buffer area towards planning/establishment of tailor-made biofuel/biomass plants.

Fine-scale multi-temporal and spatial analysis of agricultural drought in agro-ecological regions of Zimbabwe

This study utilised multi-temporal satellite data to characterize droughts across the five Agro-ecological zones (AEZ) of Zimbabwe at monthly, seasonal, sub-seasonal and annual scales (2000 to 2018) using a combination of Satellite Pour l'Observation de la Terre (SPOT) Vegetation (VGT) and Project for On-Board Autonomy-Vegetation (PROBA-V) derived Vegetation Condition Index (VCI). The characterized droughts were validated, using maize yield and meteorological station derived Standardised Precipitation Index (SPI). The study showed differential spatial and temporal expression of drought in Zimbabwe, over the 19-year study period. November had the highest frequency of drought while the first quarter of the season experienced more droughts than the second quarter. Spatially, virtually all AEZ in Zimbabwe are prone to droughts and the frequency is increasingly, with the Agro-ecological region five (southern Zimbabwe) being the worst affected. Further, the study found a strong positive relationship between the VCI and maize yield (r = 0.701, P = 0.000), demonstrating the utility of the index in drought monitoring at national level. Overall, findings from this work highlight the significance of using multi-date and national-scale analysis of agricultural drought occurrence and trends. Moreover, this work provides critical baseline information for developing drought resilience and mitigation strategies in drought-prone areas.

Detection of heartwood rot in Norway spruce trees with lidar and multi-temporal satellite data

Norway spruce pathogenic fungi causing root, butt and stem rot represent a substantial problem for the forest sector in many countries. Early detection of rot presence is important for efficient management of the forest resources but due to its nature, which does not generate evident exterior signs, it is very difficult to detect without invasive measurements. Remote sensing has been widely used to monitor forest health status in relation to many pathogens and infestations. In particular, multi-temporal remotely sensed data have shown to be useful in detecting degenerative diseases. In this study, we explored the possibility of using multi-temporal and multi-spectral satellite data to detect rot presence in Norway spruce trees in Norway. Images with four bands were acquired by the Dove satellite constellation with a spatial resolution of 3 m, ranging over three years from June 2017 to September 2019. Field data were collected in 2019–2020 by a harvester during the logging: 16163 trees were recorded, classified in terms of species and presence of rot at the stump and automatically geo-located. The analysis was carried out at individual tree crown (ITC) level, and ITCs were delineated using lidar data. ITCs were classified as healthy, infested and other species using a weighted Support Vector Machine. The results showed an underestimation of the rot presence (balanced accuracy of 56.3%, producer’s accuracies of 64.3 and 48.4% and user’s accuracies of 81.0% and 32.7% respectively for healthy and rot ITCs). The method can be used to provide a tentative map of the rot presence to guide more detailed assessments in field and harvesting activities.