Vicca Karolinoerita

Riverine plastic waste is difficult to detect and map accurately because debris ranges from small items to large floating clusters, and tropical rivers present challenging conditions, such as murky water, floating vegetation, and variable lighting. This study develops and benchmarks a deep learning method for pixel-level, multi-scale mapping of riverine waste from five-band UAV multispectral imagery. We deployed a drone equipped with a five-band multispectral sensor over the Brantas River in Surabaya, East Java, Indonesia, and introduce the Multi-Scale Hierarchical Attention Ensemble (MHAE), which combines three backbone networks across three image resolutions through learned scale- and backbone-attention weighting. MHAE was evaluated against 12 CNN-, transformer-, state-space-, and traditional-machine-learning-based baselines (including Random Forest, U-Net, UNet++, and DeepLabV3+) on the accompanying BrantasRiverWaste-UAV dataset (882 image tiles from a single-site, single-season orthomosaic covering approximately 0.54 km² of river surface, labelled as water, land, organic waste, or inorganic waste), with pairwise comparisons assessed using Wilcoxon signed-rank tests with Bonferroni correction. MHAE achieved the highest pixel-level waste detection rate among the 12 evaluated models (86.76%), with a mean intersection-over-union of 78.33% (third-highest, behind UNet++ and U-Net). This work provides an initial, single-site benchmark and reference architecture for near-real-time riverine waste monitoring, and introduces the BrantasRiverWaste-UAV as, to our knowledge, one of the first five-band multispectral UAV datasets for tropical riverine waste mapping. Balanced sampling and multi-scale attention fusion can substantially improve waste-pixel detection under severe class imbalance; because the benchmark derives from a single site and season, future work should extend evaluation across additional seasons and river systems before the approach is generalised operationally.

Water conservation (WC) is a vital ecosystem service (ES) that plays an essential role in the sustainable management of water resources and ensures ecological security. This research examines the WC capacity of the Citarum watershed in West Java Province, Indonesia, from 2010 to 2020. The specific objectives of this research are as follows: (1) to assess the 10-year temporal and spatial variations of WC using the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) and topographic index model; (2) to analyze the temporal interchange between WC and its influencing factors through the Spatial Elastic Coefficient Trajectory Model (SECTM); and (3) to determine the driving factors (DFs) affecting WC by applying Multiscale Geographically Weighted Regression (MGWR). Key findings of this study reveal significant trends in WC from 2010 to 2020: the average WC in the Citarum watershed decreased from 513.96 mm/ha to 453.64 mm/ha (11.74%), indicating a concerning decline in ES capacity. This study also noted that regions implementing comprehensive regulations dominated the study area, covering approximately 72.70% of the total area (502,295 hectares). This illustrates that the implementation of rules plays a crucial role in the WC. Lastly, the MGWR analysis found that WC in the Citarum watershed positively correlated with topography, climate, and vegetation while negatively correlating with socioeconomic factors. This indicates that WC levels are generally lower in areas with higher human activity and economic growth, highlighting the impact of anthropogenic pressures on natural resources. This framework helps stakeholders plan to ensure sustainable development in the area, as it provides valuable insights into the interactions between the ecological and socioeconomic factors affecting WC.