The complexity of measuring cropland use intensity: An empirical study

CONTEXT: Cropland intensification promises additional food supply without further expanding croplands into natural ecosystems. Cropland use intensity measures the degree of intensification by various indicators, both from input and output perspectives; all of these indicators are however not always coherent—which may cause confusion and send contradicting signals to policymakers. Few empirical studies have been conducted to relate and compare various intensity indicators. OBJECTIVE: In this study, we start with a hypothesis that cropland use intensity measured by different indicators may differ. We then test such a hypothesis based on empirical evidence in a typical multi-cropped region (i.e., Jinxian County) in Jiangxi Province—a major breadbasket in Southern China. METHODS: We focus on two widely used indicators, i.e., multi-cropping frequency (MCF) and crop growth duration (GDa), measured by a hybrid time-series remote sensing dataset which fuses MODIS and Gaofen-1 images for the year 2015. We map these two indicators independently at a 16 m spatial resolution. For each pixel, MCF takes value from 1 to 3 corresponding to single-, double- and triple- cropping; while GDa is quantified by the accumulative crop growth days within the study year. We relate the values of two indicators, summarize the descriptive statistics of GDa grouped by MCF categories, and compare MCF and GDa values by using a box-whisker chart, a bivariate map, and a set of statistical tests. RESULTS AND CONCLUSIONS: We find that a significant overlap of GDa exists between single cropping and double cropping both visually and statistically. In other words, the different cropland intensity levels measured by MCF appear to be the same if measured by GDa. The box-whisker chart shows that the GDa of single cropping in the upper quarter is substantially longer than the GDa of double cropping in the lower quarter. The non-parametric tests show that the overlapped GDa between single cropping and double cropping exists substantially, which may cause confusion. Moreover, the bivariate map shows that single cropping with shorter growth duration (i.e., consistent low intensity) and double cropping with longer growth duration (i.e., consistent high intensity) account for 44% and 28%, respectively, while the rest manifested as inconsistent which accounts for another 28% of the total cropland area. SIGNIFICANCE: These results confirm our hypothesis—the inconsistency among intensification indicators. Different measurements may convey contradicting messages for policymakers in pursuing the sustainable intensification goals, which suggests more efforts are required to understand the multidimensionality of the process of cropland intensification for distinct policy measurements.