Climate, soil and management factors drive the quantitative relationships between soil fertility and spring maize water productivity in northern China

Dryland agriculture in northern China is mainly limited by water scarcity and low soil fertility. This study thus quantified relationships between spring maize water productivity (WP) and soil fertility as well as variations in such relationships. In doing so, a total of 844 datasets were integrated by data from four 23–40 years long-term experiments (Gongzhuling, Pingliang, Shenyang and Shouyang) and 21 peer-reviewed publications. The random forest, partial least squares structural equation modeling and variance partitioning analysis were then applied to address the quantitative relationships (QRs) and relevant differences. Results showed that QRs between regions were 2.14 (Northwest China, NW) > 1.53 (North China, NC) > 0.97 kg m−3 (Northeast China, NE); 1.70, 1.41, and 1.18 kg m−3 at dry, normal and wet years; 1.09 under chemical nitrogen (N), 1.57 under NP (chemical phosphorus), 2.08 under NPK (chemical potassium) (highest), 1.29 under manure (M), 1.09 under NM, 1.39 under NPM and 0.98 kg m−3 under NPKM (lowest), respectively. Mean annual precipitation, mean annual temperature, available N, N fertilization rate and maize varieties were main factors affecting QRs, while the effects and interactions of climate, soil properties and management factors were the main causes producing differences in QRs. The potential QRs determined by boundary functions were reached up to 6.2, 5.26 and 6.12 kg m−3 in NW, at dry years and under NPK, respectively. To improve spring maize WP through enhanced soil fertility, the NW region, dry years, and NPK were parallelly optimal, while such WP increases were increased with N rate in NE and NW or planting pattern change in NC, each of them could efficiently benefit for using scarce water resources in dryland northern China and other dryland areas.