Mapping and assessing maize water-efficiency dynamics using data assimilation and the SWAP model in Northwest China

Abstract

Context

Evaluation of the water-efficiency performance of crop production is essential for achieving precise zoning management and improving yields. However, current quantitative assessments of regional crop water-efficiency show limited accuracy and lack of spatial heterogeneity.

Method

This study used ensemble Kalman filter assimilation of soil water content (SWC) and leaf area index into the Soil Water Atmosphere Plant (SWAP) model to simulate the spatiotemporal dynamics of maize water-efficiency in Wuwei of Northwest China from 2015 to 2022. An integrated water-efficiency indicator (WEI) was then developed by coupling four indicators: crop water stress index, yield loss rate, water productivity, and irrigation efficiency. The key driving factors influencing WEI variation were identified, and WEI was classified into four levels (Class I–IV), from high to low efficiency. Finally, the transitions of maize fields across classes from 2015 to 2022 were analyzed.

Results

The results show that integrating data assimilation with the SWAP model achieved high simulation accuracy for maize growth and SWC. The WEI increased by 65.79 %, rising from 0.38 to 0.63 between 2015 and 2022. The main drivers of WEI change over these eight years were irrigation, purity of maize pixels and precipitation. Class Ⅲ was relatively stable, with a self-maintenance probability of 33.3 %. Each initial class had a transition probability to Class Ⅲ above 28.3 % (average 31.0 % ± 2.0 %), showing that Class Ⅲ played a central role in system evolution. The highest transition probability was observed from Class IV to Class II (37.0 %), highlighting the effectiveness of water-saving and yield-boosting field measures.

Conclusions

This study proposed a new approach for accurately identifying water-efficiency hotspots in maize production and provided data support for guiding precise zonal management in agriculture.