Mapping Seasonal Spatiotemporal Dynamics of Alpine Grassland Forage Phosphorus Using Sentinel-2 MSI and a DRL-GP-Based Symbolic Regression Algorithm DOI Creative Commons

Jiancong Shi,

Aiwu Zhang, Juan Wang

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(21), P. 4086 - 4086

Published: Nov. 1, 2024

An accurate estimation of seasonal spatiotemporal dynamics forage phosphorus (P) content in alpine grassland is crucial for effective and livestock management. In this study, we integrated Sentinel-2 multispectral imagery (MSI) with computational hyperspectral features (CHSFs) developed a novel symbolic regression algorithm based on deep reinforcement learning genetic programming (DRL-GP) to estimate P grasslands. Using 243 field observations collected during the regreening, grass-bearing, yellowing periods 2023 from Shaliu River Basin, generated 10 CHSF images (CHSFIs) varying spectral dispersions (1–10 nm). Our results demonstrated following: (1) The DRL-GP-based model identified optimal dispersion each growing season, significantly enhancing accuracy. (2) Forage estimations using combined outperformed traditional methods. Compared original features, R2 improved by 99.5%, 57.4%, 86.2% periods, corresponding MSE reductions 84.8%, 41.5%, 75.8% MAE decreases 70.7%, 57.5%, 50.4%. Across these seasons, increased 322.2%, 68.2%, 639.8% compared MLR, 128.9%, 97.4%, 469.2% RF, 485.1%, 65.3%, 231.3% DNN. decreased 31%, 82.9%, 52.4% 39.9%, 42.4%, 31.4% 84.5%, 73.4%, 81.9% 32.6%, 67%, 44.2% 42.6%, 47.6%, 37.9% 60.2%, 50%, 56.3% (3) Proximity water system notably influenced variation, highest increases observed within 1–2 km sources. These findings provide critical insights optimizing management improving productivity.

Language: Английский

Comparison of Graph Theory and Movements Simulation Approaches in Building Ecological Networks. The Case Study of the Metropolitan Area of Reggio Calabria (Italy) DOI
Giovanni Lumia, Salvatore Praticò, Samuel A. Cushman

et al.

Lecture notes in civil engineering, Journal Year: 2025, Volume and Issue: unknown, P. 1131 - 1139

Published: Jan. 1, 2025

Language: Английский

Citations

0
Mapping Seasonal Spatiotemporal Dynamics of Alpine Grassland Forage Phosphorus Using Sentinel-2 MSI and a DRL-GP-Based Symbolic Regression Algorithm DOI Creative Commons

Jiancong Shi,

Aiwu Zhang, Juan Wang

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(21), P. 4086 - 4086

Published: Nov. 1, 2024

An accurate estimation of seasonal spatiotemporal dynamics forage phosphorus (P) content in alpine grassland is crucial for effective and livestock management. In this study, we integrated Sentinel-2 multispectral imagery (MSI) with computational hyperspectral features (CHSFs) developed a novel symbolic regression algorithm based on deep reinforcement learning genetic programming (DRL-GP) to estimate P grasslands. Using 243 field observations collected during the regreening, grass-bearing, yellowing periods 2023 from Shaliu River Basin, generated 10 CHSF images (CHSFIs) varying spectral dispersions (1–10 nm). Our results demonstrated following: (1) The DRL-GP-based model identified optimal dispersion each growing season, significantly enhancing accuracy. (2) Forage estimations using combined outperformed traditional methods. Compared original features, R2 improved by 99.5%, 57.4%, 86.2% periods, corresponding MSE reductions 84.8%, 41.5%, 75.8% MAE decreases 70.7%, 57.5%, 50.4%. Across these seasons, increased 322.2%, 68.2%, 639.8% compared MLR, 128.9%, 97.4%, 469.2% RF, 485.1%, 65.3%, 231.3% DNN. decreased 31%, 82.9%, 52.4% 39.9%, 42.4%, 31.4% 84.5%, 73.4%, 81.9% 32.6%, 67%, 44.2% 42.6%, 47.6%, 37.9% 60.2%, 50%, 56.3% (3) Proximity water system notably influenced variation, highest increases observed within 1–2 km sources. These findings provide critical insights optimizing management improving productivity.

Language: Английский

Citations

0