Agronomy, Vol. 16, Pages 236: Agricultural Irrigation Water Requirement Prediction in Arid Regions Based on the Integration of the AquaCrop-OS Model and Deep Learning: A Case Study of the Qarqan River Basin, China

Agronomy doi: 10.3390/agronomy16020236

Authors:
Fan Gao
Hairui Li
Bing He
Kun Liu
Jiacheng Zhang
Qiang Liu
Ying Li
Lu Wang

Water scarcity and ecological degradation driven by the expansion of irrigated agriculture in arid regions urgently necessitate a rigorous assessment of the combined impacts of climate change and crop-structure adjustments on irrigation water requirements (IWR). Taking the Qarqan River Basin as a case study, this study establishes an integrated framework that incorporates remote sensing (Landsat/MODIS), the AquaCrop-OS crop model, and a CNN-LSTM deep learning architecture to simulate historical IWR (2000–2024) and project future trajectories under CMIP6 climate scenarios. The results indicate that: (1) from 2000 to 2024, fruit tree area expanded from 120.3 to 320.3 km2, cotton stabilized at approximately 165.3 km2 after peaking at 187.9 km2 in 2014, wheat recovered to 113.1 km2, and maize varied between 23.7 and 85.0 km2, indicating that fruit trees have become the dominant crop type. (2) Over the same period, total basin-wide IWR increased by 91% (3.7 × 108 to 7.1 × 108 m3), with fruit trees accounting for 44–68% of this growth. Logarithmic mean Divisia index (LMDI) decomposition further shows that meteorological factors and human activities jointly drove the increase in IWR, with cultivated-area expansion and cropping-structure change contributing most, while improvements in agricultural water-use efficiency partially offset the rise. (3) Projections for 2025–2100 suggest stronger structural dominance of fruit trees and cotton; the growing share of water-intensive cash crops may further elevate irrigation pressure. Under SSP5-8.5, a 30% reduction in fruit tree area in the late century could save 4.3% of irrigation water (0.33 × 108 m3). Overall, this study provides dynamic projections and decision support for adaptive regulation of agricultural water resources in arid regions.

Source link

Fan Gao www.mdpi.com