Irrigation water consumption of irrigated cropland and its dominant factor in China from 1982 to 2015

Highlights

• A method of estimating the long-term irrigation water consumption of irrigated cropland at the regional scale was proposed.

• The spatial pattern and trend of irrigation water consumption in China from 1982 to 2015 was quantified.

• Highly expanded irrigated farmland dominates the significant increase in total annual irrigation water consumption of irrigated cropland in China (67.2%±5.6%, 64.9%±4.7%).

• The increased amplitude of annual total blue water consumption in the arid zone was higher than that in the humid area.

Abstract

Quantifying and mapping regional and global irrigation water consumption have attracted considerable attention from researchers. The existing models cannot accurately estimate the long-term sequence of irrigation water consumption at the regional scale. Despite being among the countries with the largest irrigated areas, China has not been specifically evaluated in terms of high-resolution long-term irrigation water consumption. In this paper, an irrigated cropland water model incorporating irrigated cropland mapping, the soil water balance equation, irrigated crop phenology, and partitioning evapotranspiration products was developed to fill the knowledge gap. We found that (1) the mean annual total irrigation water consumption of irrigated cropland in China during 1982–2015 was approximately 179.43 ± 24.60 km³. (2) The continued expansion of irrigated farmland dominated (67.2% ± 5.6%) the significant increase in annual total irrigation water consumption (slope=2.03 km³ yr⁻¹, p<0.01), followed by changes in crop management practices (16.7% ± 6.1%), and the impacts of interannual climatic variability (represented by precipitation and potential evapotranspiration in this research) were weak. (3) The annual total irrigation water consumption in the arid zone increased quicker than that in wet regions.

Introduction

The water that meets the requirements for crop production can be separated into precipitation (or ‘green water’) and irrigation water (or ‘blue water’) (Falkenmark and Rockstrom 2006; Falkenmark 2007; Romaguera et al., 2014). The irrigation water from rivers, lakes and aquifers is the largest portion of human freshwater consumption, accounting for more than 90% of the total global freshwater consumption (Siebert and Döll 2010; Hoekstra and Mekonnen 2012). With the rapid growth of the global population and the increasing demand for food over the past decades, irrigation water consumption has increased with the expansion of irrigated cropland (Siebert et al., 2015). The increase in human water consumption is expected to intensify water scarcity in many countries (Oki and Kanae 2006; Hayashi et al., 2013; Wada et al., 2016). Therefore, assessing irrigation water consumption and its implications for the sustainable use of freshwater, related food security issues and efficient management of water resources is critical, especially in water-limited areas (Velpuri and Senay 2017; Huang et al., 2019).

Quantifying and mapping regional and global irrigation water consumption have captured considerable attention (Siebert and Döll 2010; Mekonnen and Hoekstra 2011; Hoekstra and Mekonnen 2012; Romaguera et al., 2012; Romaguera et al., 2014; Hoogeveen et al., 2015; Chen et al., 2019). Initial estimates of irrigation water consumption (Postel 1998; Rockstrom and Gordon 2001) were fairly crude, based on compilations of various static and possibly inconsistent vegetation maps, average crop productivities, yields, and ET rates. Later, global irrigation water consumption estimation methods focused on macroscale hydrological models and crop models (de Fraiture 2007; Hanasaki et al., 2008; Rost et al., 2008; Liu and Yang 2010; Siebert and Döll 2010; Huang et al., 2019). For instance, Rost et al. (2008) reported that global irrigation water consumption was 1364 km³ by using the dynamic global vegetation and water balance model from 1971 to 2000. Siebert and Döll (2010) proposed a global crop water model to calculate irrigation water use, and the estimated irrigation water use was 1180 km³ during 1998–2002. Chen et al. (2019) proposed the framework of observationally constrained cropland water consumption calculations to estimate the global irrigation water consumption, which was 874 km³ (in 2005). Recently, Huang et al. (2019) predicted future global agricultural irrigation water consumption by a global change assessment model and its hydrology module. However, these models need to be further improved before they can be applied at the regional scale, as these models should incorporate high-temporal and high spatial climate variables, as well as improve irrigation and rain-fed classification resolutions.

At the regional scale, Romaguera et al. (2012) estimated the blue water consumption (irrigation water consumption) of irrigated cropland as the difference between the two evapotranspiration (ET) products that were calibrated. One ET product is driven by precipitation and only represents precipitation consumption (e.g., Global Land Data Assimilation System Noah (GLDAS)), while the other ET product considers irrigation practices, including irrigation water consumption and precipitation consumption (e.g., energy balance-based ET). The ET datasets will be agreed upon after reliable calibration, with their difference being blue water, and the value of blue water is comparable to another ecosystem ET. Romaguera et al. (2014) applied this method in irrigated agricultural areas to estimate monthly irrigation water consumption based on the GLDAS ET (0.25°) product and surface energy balance system ET product (0.1°) over 2001–2012. However, the mismatched scale and the lack of reliable flux data make the results more uncertain. Velpuri and Senay (2017) solved that problem by combining two identical 1 km actual ET datasets, one obtained from a root zone water balance model that captured precipitation water use only (Senay 2008) and another from an energy balance model (Gabriel et al., 2013), as well as 21 reliable FLUXNET sites over 2001–2015. However, the calibration process is extremely sensitive to input data (Romaguera et al., 2012). Many reliable ET observations in irrigated cropland flux tower sites used for calibration are also expensive and difficult to obtain, as well as only available in data-rich areas (e.g., the United States, Europe), which would limit the application of this method in areas where flux data are scarce. Hoekstra (2019) proposed a generic and physically based method to estimate irrigation water consumption but did not apply it to any research area, resulting in questions of whether the simulation works well or not.

In addition, the existing research time period at the regional scale is short, and it does not characterize the long-term variability in irrigation water consumption. This raises the question of how to develop a consistent framework to estimate the long-term irrigation water consumption of irrigated cropland and make it comparable to natural ecosystem ET. As one of the countries with the largest irrigated area, China must be able to feed the world's largest population. Understanding the quantity and variability in irrigation water consumption of irrigated cropland helps guide water resource-related policies. However, the specific evaluation of high-resolution long-term irrigation water consumption for China has not been reported. Thus, the objective of this paper is to (1) develop an easy-to-implement methodology for an irrigated cropland water model for irrigation water consumption estimation, (2) capture the high-resolution long-term temporal and spatial patterns of irrigation water consumption over China during 1982–2015, and (3) quantify the relative importance of factors affecting irrigation water consumption variation.