In Indonesia, flooding is one of the natural hazards that often occurs during the rainy season. Surface runoff coefficient values are an essential indicator of the supply of regional water resources. The smaller the surface runoff value, the greater the water storage in the ground, and the smaller surface was running water. This study analyses the spatial and temporal distribution of the estimated surface runoff caused by land use/land cover changes in the upstream Citarum watershed. The study area is located in the upstream Citarum watershed, West Java, Indonesia. The site has a long history of flooding and various complex environmental problems. The geographic Information System method was used as a tool in analyzing the spatially and temporally. The research result shows that there has been a change in land cover in several periods of the year in the Citarum upstream watershed. The occurrence of the LULC phenomenon positively affects the surface runoff coefficient. The increasing area of Built land and plantation in the Citarum upstream watershed will further increase the surface runoff coefficient and, in the end, will potentially increase the surface runoff and contribute to flooding in the Bandung basin. This study results can be used to provide input in determining the direction and policies for watershed management, taking into account the varying characteristics of each subwatershed.

Indonesia is one of the countries in the world that is frequently affected by floods. Flood disasters can have various negative impacts and therefore need to be analyzed to determine prevention and mitigation measures. This study examined land cover change, flood detection, and flood distribution using multitemporal Sentinel-1 and Landsat-8 satellite imagery in the Barito watershed. A combination of change detection and the application of the Otsu algorithm was used to detect floodplains from Sentinel-1 imagery. Land use/land cover (LULC) changes are detected using a combination of change detection and machine learning in the form of a random forest algorithm. The overlay technique was used to analyze the distribution of floodplains. In this study, the floodplain in the study area was mapped to 109,623 ha. The change detection method detects a decrease in the areas of primary forest, secondary forest, fields, rice fields, shrubs and ponds, respectively, by 13,020 ha, 116,235 ha, 259 ha, 146,696 ha, 47,308 ha, and 9,601 ha. Settlements, bare land, plantations and water bodies increase by 14,879 ha, 64,830 ha, 218,916 ha, and 34,768 ha, respectively. Flooding was mainly found in the classes of rice fields, water bodies and primary forests.

In this study, we proposed an automatic water extraction index (AWEI) threshold improvement model that can be used to detect lake surface water based on optical remote sensing data. An annual Landsat 8 mosaic was created using the Google Earth Engine (GEE) platform to obtain cloud-free satellite image data. The challenge of this study was to determine the threshold value, which is essential to show the boundary between water and nonwater. The AWEI was selected for the study to address this challenge. The AWEI approach was developed by adding a threshold water value based on the split-based approach (SBA) calculation analysis for Landsat 8 satellite images. The SBA was used to determine local threshold variations in data scenes that were used to classify water and nonwater. The class threshold between water and nonwater in each selected subscene image can be determined based on the calculation of class intervals generated by geostatistical analysis, initially referred to as smart quantiles. It was used to determine the class separation between water and nonwater in the resulting subscene images. The objectives of this study were (a) to increase the accuracy of automatic lake surface water detection by improvising the determination of threshold values based on analysis and calculations using the SBA and (b) to conduct a test case study of AWEI threshold improvement on several lakes’ surface water, which has a variety of different or heterogeneous characteristics. The results show that the threshold value obtained based on the smart quantile calculation from the natural break approach (AWEI ≥ −0.23) gave an overall accuracy of close to 100%. Those results were better than the normal threshold (AWEI ≥ 0.00), with an overall accuracy of 98%. It shows that there has been an increase of 2% in the accuracy based on the confusion matrix calculation. In addition to that, the results obtained when classifying water and nonwater classes for the different national priority lakes in Indonesia vary in overall accuracy from 94% to 100%.

No abstract

Improvement in the accuracy of the postclassification of land use and land cover (LULC) is important to fulfil the need for the rapid mapping of LULC that can describe the changing conditions of phenomena resulting from interactions between humans and the environment. This study proposes the majority of segment-based filtering (MaSegFil) as an approach that can be used for spatial filters of supervised digital classification results. Three digital classification approaches, namely, maximum likelihood (ML), random forest (RF), and the support vector machine (SVM), were applied to test the improvement in the accuracy of LULC postclassification using the MaSegFil approach, based on annual cloud-free Landsat 8 satellite imagery data from 2019. The results of the accuracy assessment for the ML, RF, and SVM classifications before implementing the MaSegFil approach were 73.6%, 77.7%, and 77.5%, respectively. In addition, after using this approach, which was able to reduce pixel noise from the results of the ML, RF, and SVM classifications, there were increases in the accuracy of 81.7%, 85.2%, and 84.3%, respectively. Furthermore, the method that has the best accuracy RF classifier was applied to several national priority watershed locations in Indonesia. The results show that the use of the MaSegFil approach implemented on these watersheds to classify LULC had a variation in overall accuracy ranging from 83.28% to 89.76% and an accuracy improvement of 6.41% to 15.83%.

The study area is a tropical environment located in the Batanghari watershed, Sumatra, Indonesia. The existence of the environmental problems and damages in the study area can be identified based on land degradation. It can be interpreted as a complex process and is influenced by human activities, climate change, and natural events. This study proposes the latest Geospatial Artificial Intelligence (Geo-AI) model using multi-sources geospatial data that is specifically used to address challenges and phenomena related to the identification of land degradation in the study area. The novelty of this study is that it is the first time to integrate the 6 (six) main variables of multi-source geospatial data - Topographical, Biophysical, Bioclimatic, Geo-environmental, Global human modification, and Accessibility - in predicting potential land degradation in the tropical environment, such as Indonesia. Machine learning-based prediction Support Vector Machine (SVM), Minimum Distance (MD), Classification and Regression Trees (CART), Gradient Tree Boost (GTB), Naïve Bayes (NB), Random Forest (RF) algorithms were used to predict and to map land degradation in the study area. The overall accuracy of the results of comparison and evaluation of machine learning-based predictions on the RF, CART, GTB, SVM, NB, and MD in the study area are 86.2%, 85.8%, 81.2%, 52.8%, 36.3%, and 34.5%, respectively. Therefore, the study concluded that the RF, CART, and GTB algorithms are proposed to be applied to produce land degradation map in the study area.

This study proposes a new model for flood inundation modeling using the Raster-based Probability Flood Inundation Model (RProFIM) approach. The flood modeling was carried out based on the landuse/landcover (LULC) change scenario and the difference in return periods in the study area. The aims of this study were: a) to estimate the volume of discharge in the LULC change scenario between 1990 and 2050 and the difference in return periods between 2 and 100 years; b) to create and produce flood inundation maps using the RProFIM approach; and c) to analyse flood damage assessment based on the results provided by overlaying LULC data with flood inundation maps. In general, the flood probability modeling generated by RProFIM provided the same pattern and conditions it was shown by reference data. The results found several potential areas in Citarum watershed, West Java-Indonesia which are likely to be flooded based on the RProFIM approach in the Districts of Margaasih, Kutawaringin, Margahayu, Katapang, Dayeuhkolot, and Baleendah. Flood damage assessment by the flood scenario with a return period of 2–100 years and changes in LULC in the range of 1990–2050, shows that the largest estimated loss based on market value is on built-up land and agriculture. This research is expected to be used as one of the considerations in managing environmental problems in overcoming flooding in the study area.

No abstract