Discover geoscience最新文献

Geospatial technologies now allow routine observation of lakes and wetlands across large areas, but turning those observations into timely, actionable insight still requires scalable computing. Google Earth Engine (GEE) provides a web-based platform that brings multi-sensor Remote Sensing (RS) archives and parallel processing together in one environment. This review synthesizes how artificial intelligence (AI), machine learning (ML) and deep learning (DL) have been paired with GEE to map and monitor surface water quantity and quality. We summarize recent methods, compare model families commonly used on GEE, and discuss frequent processing pitfalls. To ground the review, we include a case study of three Nebraska lakes (2022-2023) that demonstrates month-to-month tracking of water extent and indicators of water quality. The results demonstrated the effectiveness of GEE in providing timely and accurate insights for surface water monitoring and assessment while also revealing current limitations and opportunities for improvement. Overall, we find that coupling AI methods with GEE can strengthen operational surface water assessment and inform decision-making under increasing environmental pressures.

Supplementary information: The online version contains supplementary material available at 10.1007/s44288-025-00255-x.

Diffuse pollution, globally affecting water quality by delivery of sediment, nutrients, pathogens and agro-chemicals from farmland, often has dominant flowpaths connecting to discrete channel delivery points, where field-edge mitigation can be optimally targeted. Accurate representation of field convergent flow paths (CFPs) can inform decisions on riparian mitigation planning. For three fields in Wexford, Ireland, we combined literature, catchment data, field-survey and spatial data methods to derive sediment and P exports (7.4-18.7 tonnes sediment/year and 0.9-6.9 kgP/year), runoff areas and watercourse delivery points (one to six CFPs per field). We moderated exports according to the ratio effective riparian buffer area: CFP contributing area and compared 3 mitigation levels. Low buffer to CFP area ratios highlighted limitations of narrow buffers for larger CFPs. Linear grass buffers (2 m, level 1) were predicted to retain 2-17% of sediment and 1-6% total P exports. Level 2, 5 m buffers targeting CFP delivery points to watercourses retained 4-38% of the sediment and 2-15% total P and improved cost-effectiveness two- to three- fold relative to level 1 (20-1761 Euros/tonne sediment and 650-5114 Euros/kgP for level 2). Level 3 scenarios (sediment traps and in-ditch filters; 49% and 33% retention of field sediment and P losses, respectively) improved cost-effectiveness (50-145 Euros/tonne sediment and 108-1498 Euros/kgP). Mitigation cost-effectiveness best informs policy and planning and landowner decisions by including surface runoff behaviour utilising spatial soil and topographic data, accompanied by walk-over ground truthing.

Supplementary information: The online version contains supplementary material available at 10.1007/s44288-025-00125-6.

Climate change is likely to exacerbate land to water phosphorus (P) transfers, causing a degradation of water quality in freshwater bodies in Northwestern Europe. Planning for mitigation measures requires an understanding of P loss processes under such conditions. This study assesses how climate induced changes to hydrology will likely influence the P transfer continuum in six contrasting river catchments using Irish national observatories as exemplars. Changes or stability of total P (TP) and total reactive P (TRP) transfer processes were estimated using far-future scenarios (RCP4.5 and RCP8.5) of modelled river discharge under climate change and observed links between hydrological regimes (baseflow and flashiness indices) and transfer processes (mobilisation and delivery indices). While there were no differences in P mobilisation between RCP4.5 and RCP8.5, both mobilisation and delivery were higher for TP. Comparing data from 2080 (2070-2099) with 2020 (2010-2039), suggests that P mobilisation is expected to be relatively stable for the different catchments. While P delivery is highest in hydrologically flashy catchments, the largest increases were in groundwater-fed catchments in RCP8.5 (+ 22% for TRP and + 24% for TP). The inter-annual variability of P delivery in the groundwater-fed catchments is also expected to increase. Since the magnitude of a P source may not fully define its mobility, and hydrological connections of mobilisation areas are expected to increase, we recommend identifying critical mobilisation areas to target future mitigation strategies. These are hydrologically connected areas where controls such as soil/bedrock chemistry, biological activity and hydrological processes are favourable for P mobilisation.