Coastal regions are one of the most vulnerable areas to the effects of global warming, which is accompanied by an increase in mean sea level and changing shoreline configurations. In Indonesia, the socioeconomic importance of coastal regions where the most populated cities are located is high. However, shoreline changes in Indonesia are relatively understudied. In particular, detailed monitoring with remote sensing data is lacking despite the abundance of datasets and the availability of easily accessible cloud computing platforms such as the Google Earth Engine that are able to perform multi-temporal and multi-sensor mapping. Our study aimed to assess shoreline changes in East Java Province Indonesia from 2000 to 2019 using variables derived from a multi-sensor combination of optical remote sensing data (Landsat-7 ETM and Landsat-8 OLI) and radar data (ALOS Palsar and Sentinel-1 data). Random forest and GMO maximum entropy (GMO-Maxent) accuracy was assessed for the classification of land and water, and the land polygons from the best algorithm were used for deriving shorelines. In addition, shoreline changes were quantified using Digital Shoreline Analysis System (DSAS). Our results showed that coastal accretion is more profound than coastal erosion in East Java Province with average rates of change of +4.12 (end point rate, EPR) and +4.26 m/year (weighted linear rate, WLR) from 2000 to 2019. In addition, some parts of the shorelines in the study area experienced massive changes, especially in the deltas of the Bengawan Solo and Brantas/Porong river with rates of change (EPR) between −87.44 to +89.65 and −18.98 to +111.75 m/year, respectively. In the study areas, coastal erosion happened mostly in the mangrove and aquaculture areas, while the accreted areas were used mostly as aquaculture and mangrove areas. The massive shoreline changes in this area require better monitoring to mitigate the potential risks of coastal erosion and to better manage coastal sedimentation.

Machine learning has been employed for various mapping and modeling tasks using input variables from different sources of remote sensing data. For feature selection involving high- spatial and spectral dimensionality data, various methods have been developed and incorporated into the machine learning framework to ensure an efficient and optimal computational process. This research aims to assess the accuracy of various feature selection and machine learning methods for estimating forest height using AISA (airborne imaging spectrometer for applications) hyperspectral bands (479 bands) and airborne light detection and ranging (lidar) height metrics (36 metrics), alone and combined. Feature selection and dimensionality reduction using Boruta (BO), principal component analysis (PCA), simulated annealing (SA), and genetic algorithm (GA) in combination with machine learning algorithms such as multivariate adaptive regression spline (MARS), extra trees (ET), support vector regression (SVR) with radial basis function, and extreme gradient boosting (XGB) with trees (XGbtree and XGBdart) and linear (XGBlin) classifiers were evaluated. The results demonstrated that the combinations of BO-XGBdart and BO-SVR delivered the best model performance for estimating tropical forest height by combining lidar and hyperspectral data, with R2 = 0.53 and RMSE = 1.7 m (18.4% of nRMSE and 0.046 m of bias) for BO-XGBdart and R2 = 0.51 and RMSE = 1.8 m (15.8% of nRMSE and −0.244 m of bias) for BO-SVR. Our study also demonstrated the effectiveness of BO for variables selection; it could reduce 95% of the data to select the 29 most important variables from the initial 516 variables from lidar metrics and hyperspectral data.

Crop intensity information describes the productivity and the sustainability of agricultural land. This information can be used to determine which agricultural lands should be prioritized for intensification or protection. Time-series data from remote sensing can be used to derive the crop intensity information; however, this application is limited when using medium to coarse resolution data. This study aims to use 3.7 m-PlanetScope™ Dove constellation data, which provides daily observations, to map crop intensity information for agricultural land in Magelang District, Indonesia. Two-stage histogram matching, before and after the monthly median composites, is used to normalize the PlanetScope data and to generate monthly data to map crop intensity information. Several methods including Time-Weighted Dynamic Time Warping (TWDTW) and the machine-learning algorithms: Random Forest (RF), Extremely Randomized Trees (ET), and Extreme Gradient Boosting (XGB) are employed in this study, and the results are validated using field survey data. Our results show that XGB generated the highest overall accuracy (OA) (95 ± 4%), followed by RF (92 ± 5%), ET (87 ± 6%), and TWDTW (81 ± 8%), for mapping four-classes of cropping intensity, with the near-infrared (NIR) band being the most important variable for identifying cropping intensity. This study demonstrates the potential of PlanetScope data for the production of cropping intensity maps at detailed resolutions.

The rise of Google Earth Engine, a cloud computing platform for spatial data, has unlocked seamless integration for multi-sensor and multi-temporal analysis, which is useful for the identification of land-cover classes based on their temporal characteristics. Our study aims to employ temporal patterns from monthly-median Sentinel-1 (S1) C-band synthetic aperture radar data and cloud-filled monthly spectral indices, i.e., Normalized Difference Vegetation Index (NDVI), Modified Normalized Difference Water Index (MNDWI), and Normalized Difference Built-up Index (NDBI), from Landsat 8 (L8) OLI for mapping rice cropland areas in the northern part of Central Java Province, Indonesia. The harmonic function was used to fill the cloud and cloud-masked values in the spectral indices from Landsat 8 data, and smile Random Forests (RF) and Classification And Regression Trees (CART) algorithms were used to map rice cropland areas using a combination of monthly S1 and monthly harmonic L8 spectral indices. An additional terrain variable, Terrain Roughness Index (TRI) from the SRTM dataset, was also included in the analysis. Our results demonstrated that RF models with 50 (RF50) and 80 (RF80) trees yielded better accuracy for mapping the extent of paddy fields, with user accuracies of 85.65% (RF50) and 85.75% (RF80), and producer accuracies of 91.63% (RF80) and 93.48% (RF50) (overall accuracies of 92.10% (RF80) and 92.47% (RF50)), respectively, while CART yielded a user accuracy of only 84.83% and a producer accuracy of 80.86%. The model variable importance in both RF50 and RF80 models showed that vertical transmit and horizontal receive (VH) polarization and harmonic-fitted NDVI were identified as the top five important variables, and the variables representing February, April, June, and December contributed more to the RF model. The detection of VH and NDVI as the top variables which contributed up to 51% of the Random Forest model indicated the importance of the multi-sensor combination for the identification of paddy fields.

No abstract