Under global climate change, we can adjust landscape patterns to mitigate the effects of climate factors on crop yield. Field and landscape-scale research suggests that increased land cover diversity and non-crop habitats can improve crop yield, but little is known on the effects of patch area and shape of one type of land use (i.e., class-level) and spatial patterns (aggregation, fragmentation, and heterogeneity) of surrounding multiple land-use classes (i.e. landscape-level) on crop yield. We evaluated the relationship between 375 landscape metrics at both class- and landscape-levels, 120 climate variables, and the crop yields of maize, rice, and wheat in China from years 2000–2015 at national and sub-climatic zone scales. Our results suggest that both class- and landscape-level patterns were the main drivers, contributing averagely 12% of the variations in crop yields. Specifically at national scale, crop yield in cropland with forest cover (>27%) was increased by 169% than where without forest cover. Shape complexity in forest and cropland patches, aggregation degree, and spatial complexity of land cover types were more closely associated with crop yield (accounting for 2–73% of the variations in crop yield at sub-climatic zones) than landscape diversity (<2%), mainly owing to their mediating effects on the relationship between climate factors and crop yield. Not all non-crop habitats are beneficial to crop yield, and forest patches with a suitable area and shape complexity perform best. Our findings suggest that the regulation of both class- and landscape-level patterns may be an alternative way to improve crop yield, rather than only enhancing land-cover diversity or increasing the area of non-crop habitats.
Read full abstract