A comprehensive analysis of water productivity in natural vegetation and various crops coexistent agro-ecosystems

Highlights

• Equivalent water productivity (WP) was proposed and proved to be useful in WP assessment.

• WP is scale dependent due to water reuse through shallow groundwater systems.

• Natural vegetation should not be ignored when considering water productivity of a whole region.

• Water reuse reduces WP compared with direct use in salt affected areas.

• Efficient irrigation systems are needed for WP improvement in salt affected areas.

Abstract

Water productivity (WP) expressed as the yield produced per unit volume of water is an important indicator of water use in arid and semi-arid areas. Due to complex plant cover and hydrological processes, the quantification and assessment of WP are usually difficult to determine, especially at the regional scale. In this study, an arid irrigated agro-ecosystem in the upper Yellow River basin was selected as a case study area. Based on field observation and model simulation results, the WP of irrigation water (WP_I), water applied (WP_(I+P)) and evapotranspiration (WP_ET) were calculated. Equivalent water productivity (EWP) of irrigation water (EWP_I), water applied (EWP_(I+P)) and evapotranspiration (EWP_ET) were proposed and calculated to unify the disparate WP for various crops and natural vegetation. Results showed WP_I and WP_(I+P) decreased with the increase of water application for all plants except watermelon, indicating supplemental irrigation to watermelon is urgent to improve its production and WP_(I+P). The spatially averaged WP_ET (kg m⁻³) was 2.47 for maize, 0.80 for sunflower, 12.3 for watermelon, 1.39 for wheat and 0.65 for natural vegetation. WP_ET for natural vegetation was usually lower in this salt stressed area compared with other water stressed areas. The EWP revealed the rank order of WP for different crops and natural vegetation: watermelon > wheat > maize > sunflower for EWP_I; natural vegetation > wheat > watermelon > maize > sunflower for EWP_(I+P); and wheat > sunflower > watermelon > maize > natural vegetation for EWP_ET. The relationship between EWP_(I+P) and EWP_ET was scale dependent due to the water reuse phenomena among different land cover types and the canals through the shallow groundwater system. Ignoring natural vegetation will result in considerable bias in the estimation of the regional scale water productivity (16 % in this study). WP improvement strategies such as transferring irrigation water from less productive (sunflower) areas to productive (vegetable and natural vegetation) areas, reducing bare soil evaporation and constructing a timely and accurate irrigation-drainage system were provided.

Introduction

About 40 % of the land and 35 % of the world’s population is under arid and semi-arid climatic conditions (El-Beltagy and Madkour, 2012), where water scarcity threatens food security, environmental health and economic development. In these areas, water productivity (WP), expressed as the yield or benefit derived from the use of water, has gained prominence (Molden and Sakthivadivel, 1999; Hamdy et al., 2003; Playán and Mateos, 2006; Pereira et al., 2012; Kang et al., 2017). Improving water productivity in arid irrigated agro-ecosystems is considered as the key point for saving water and sustaining agricultural production and ecosystems.

Due to land cover fragmentation and the spatial heterogeneity of land quality (such as soil conditions and irrigation accessibility), various kinds of crops usually exist in an irrigation system (Ren et al., 2019b). The yields of different crops are often not comparable, e.g. the yield of vegetables can reach about 80 tons per hectare while oil crops are only several tons (Ren et al., 2016; Zhang et al., 2017). The market prices of different crops, especially vegetables and fruits vary greatly during years. In addition, natural vegetation patches around croplands are often viewed inherently indispensable, as viable land resources for producing forage or biofuel (McLaughlin, 2011; Feng et al., 2015), draining excess water and salt (Ren et al., 2017), reducing runoff and pollutants (Liu et al., 2017b) and providing biodiversity refuges and habitat corridors (Tscharntke et al., 2002). Thus, it is not easy to compare the water productivity of different crops and natural vegetation and to quantify the productivity of a whole agro-ecosystem.

Because of the aforementioned difficulties, traditional water productivity assessment for an agro-ecosystem is usually concentrated on the comparison of WP of the same crop (Singh et al., 2006; Jiang et al., 2015; Sun et al., 2017). For the whole system, previous studies typically used a simple accumulation of the yields of different crops for calculating regional scale water productivity (Ren et al., 2019a; Huang and Li, 2010; Cao et al., 2012). However, these methods have limitations in showing real spatial trends of field WP, giving reasonable quantification of the regional WP and comparing WP between regions. In arid irrigated agro-ecosystems, not only is crop growth highly reliant on irrigation water, but also the prosperity of natural vegetation is directly or indirectly dependent on it, e.g. the reuse of field percolation and canal seepage through capillary rise and the shallow groundwater system (Ayars et al., 2006; Ren et al., 2018b). Due to the intense hydrological connectivity between the cropland and the natural land, water productivity assessment for an agro-ecosystem should consider not only the crops, but also the natural vegetation. However, in traditional water productivity assessment, natural vegetation in irrigation districts is usually ignored (Singh et al., 2006; Jiang et al., 2015; Xue et al., 2018). Above all, a reasonable criterion to unify the WPs of different crops and natural vegetation and integrate them for a whole region is necessary for water resource management.

Quantifying and improving water productivity should be based on a clear understanding of the hydrological processes (Molden et al., 2007) and the cause-effect relationships between hydrological variables (e.g. evaporation, transpiration, percolation and capillary rise) and biophysical variables (e.g. dry matter and grain yields) (Singh et al., 2006). In addition, water productivity gains can only be accomplished by optimizing it both at the field scale and the regional scale (Bouman, 2007). In irrigated agro-ecosystems with shallow groundwater table depth, the hydrological processes are very complex with frequent lateral groundwater exchange among different crop fields, canals and drainage ditches (Ren et al., 2016, 2019b; Konukcu et al., 2006). The shallow groundwater system plays an important role in the redistribution and reuse of the percolation or seepage water (Ren et al., 2018b). Thus, the quantification of the required hydrological variables is difficult. The spatial distribution of crop yields is also difficult to obtain. Field experiments and regional observations are fundamental methods to quantify water productivity. However, they are usually time-consuming and costly. Various kinds of simulation models have been developed, and they are cost-effective to quantify hydrological processes with appropriate field observations. Therefore, the use of simulation models from field to regional scales should be an attractive approach to overcome the aforementioned difficulties for water productivity assessment (Droogers and Kite, 1999; Singh et al., 2006; Xu et al., 2019).

For these reasons, an arid irrigated agro-ecosystem located in the upper Yellow River basin was taken as a case study area to explore an appropriate approach for water productivity assessment. The objectives of this study were to (1) propose a reasonable criterion to unify the water productivity of different crops and natural vegetation; and (2) present a comprehensive analysis of water productivity in irrigated agroecosystems from the field scale to the regional scale.