A model for estimating Ag-MAR flooding duration based on crop tolerance, root depth, and soil texture data

Highlights

• Agricultural managed aquifer recharge duration was estimated using a simple model.

• Model is based on plant tolerance, soil aeration, and soil hydraulic parameters.

• Potential agriculture recharge for each USDA soil texture class was estimated.

• Clay soils are usually unsuitable for agricultural recharge regardless of crop type.

• Model performance is excellent for well-defined soil and crop parameters.

Abstract

Agricultural Managed Aquifer Recharge (Ag-MAR) is an emerging MAR technique that uses agricultural fields as percolation basins to recharge the underlying aquifers. Ag-MAR can be a beneficial solution for storing excess surface water, however, if not managed properly it can potentially harm the soil and crops planted on the field at the time of recharge, ultimately leading to yield loss. Root zone residence time (RZRT), defined as the duration that the root-zone can remain saturated (or nearly saturated) during Ag-MAR without crop damage, is a key factor in Ag-MAR since extended periods of saturation in the root-zone can damage crops. Here we propose a simple RZRT model for estimating a safe Ag-MAR flooding duration based on hydraulic parameters deduced from soil texture, crop tolerance to saturation, effective root depth, and critical soil water content, which is the point where soil re-aeration occurs during drainage. We tested the model with different hydraulic parameter sets and compared the results to observed data and HYDRUS simulations. Using fitted and unfitted hydraulic parameters the average error of the predicted Ag-MAR flooding duration was less than 5 h, and up to a few days, respectively. Consequently, for crops with low flooding-tolerance, the model should be used with caution, but for more tolerant crops, the model provides reasonable predictions. The model also provides a first approximation of the possible amount of water that can be applied during an Ag-MAR event. Based on the RZRT model, we evaluated the Ag-MAR potential of various crops and effective root depths for each of the USDA soil texture classes. A spreadsheet containing the RZRT model including hydraulic parameters, and crop properties is publicly available and can be used as a learning tool or to estimate Ag-MAR flooding duration for different soils. The proposed model can be easily integrated into Ag-MAR assessment tools.

Introduction

Agricultural managed aquifer recharge (Ag-MAR) is a recharge technique for groundwater replenishment, in which farmland is flooded during the winter using excess surface water in order to recharge the underlying aquifer (Bachand et al., 2014, Dahlke et al., 2018b, Kocis and Dahlke, 2017). In California, for example, Ag-MAR is currently being implemented as part of the efforts to mitigate California’s chronic groundwater overdraft (Faunt et al., 2016, Harter, 2015; SGMA, https://water.ca.gov/programs/groundwater-management/sgma-groundwater-management).

Ag-MAR poses several risks for agricultural fields and groundwater that may influence its future adoption. This includes crop tolerance to flooding, soil aeration, biogeochemical transformations, long-term impact on soil texture, leaching of pesticides and fertilizers to groundwater, and potential greenhouse gas emissions. Some of these issues have been addressed in recent studies of Ag-MAR, including soil suitability guidelines (O’Geen et al., 2015), nitrate leaching to groundwater (Bachand et al., 2014, Bastani and Harter, 2019, Waterhouse et al., 2020), crop suitability (Dahlke et al., 2018a) and soil aeration (Bachand et al., 2019, Ganot and Dahlke, 2021). In the current study, we focused solely on the question of “how long can water be applied for Ag-MAR with minimal crop damage?”, while ignoring some of the above-mentioned challenges involving Ag-MAR implementation.

Preferably, Ag-MAR flooding is done during fallow or dormant periods, when crop damage is potentially minimal, so agricultural lands can serve as spreading basins for groundwater recharge. Root zone residence time (RZRT) is defined as the duration that the root-zone can remain saturated (or nearly saturated) during Ag-MAR without crop damage (O’Geen et al., 2015). RZRT is a crucial factor in Ag-MAR, as long periods of saturated conditions in the root-zone can damage crops due to oxygen deficiency (hypoxia) or complete depletion of oxygen (anoxia), which ultimately may result in yield loss (Kozlowski, 1997). However, flood tolerance among crops varies considerably due to biotic and abiotic conditions (Schaffer et al., 1992), therefore only appropriate crops under specific conditions may be suitable for Ag-MAR application. For example, Dokoozlian et al. (1987) have found that grapevine during dormancy can be flooded for 32 days (with an average daily recharge of 8 cm) each year without yield loss. Dahlke et al. (2018a) recently investigated the effect of different Ag-MAR flooding schemes (max. average daily recharge of 25 cm) on established alfalfa fields. Results suggest a minimal effect on yield when dormant alfalfa fields on highly permeable soils are subject to winter flooding. On the other hand, some crops are sensitive even to short-period flooding. Kiwi vines for example, are highly sensitive to root anoxia with reported yield lost and vines death due to extreme rainfalls and/or shallow groundwater levels (Smith and Buwalda, 1994). In a study on peach trees, flood cycles of 12 h per day with 5 cm ponding, applied for two months, resulted in branches with lower diameter and length growth, as well as smaller, low-quality, fruits, compared to the control trees (Insausti and Gorjón, 2013). The above examples demonstrate the need for an RZRT planning tool that can estimate Ag-MAR flood duration with minimal crop damage.

Usually, when Ag-MAR water application starts, aeration of the root-zone will be quickly suppressed by a water-layer covering the soil surface, as it prevents oxygen transport to the root-zone in the gas phase. When water application ceases, re-aeration of the root-zone will depend on the soil’s drainage rate that controls the formation of connected air pores between the root-zone and atmosphere (Fig. 1a). Hence, proper estimation of the planned flood duration during Ag-MAR requires prior knowledge of both crop characteristics and soil texture.

Only a few attempts for estimating RZRT during Ag-MAR were made, as Ag-MAR is a relatively new MAR technique. O’Geen et al. (2015) used a fuzzy logic approach to rate the RZRT during Ag-MAR, based on the harmonic mean of the saturated hydraulic conductivity (K_s) of all soil horizons, soil drainage class, and shrink-swell properties. Their RZRT rating was combined with other factors generating a Soil Agricultural Groundwater Banking Index (SAGBI, https://casoilresource.lawr.ucdavis.edu/sagbi/). Flores-Lopez et al. (2019) proposed a root-zone model that includes crop type, soil properties, and recharge suitability (based on SAGBI) to estimate water application, flooding duration, and the interval between water applications. Their model was integrated with a Groundwater Recharge Assessment Tool (GRAT; https://gratviewer.earthgenome.org/) to optimize Ag-MAR water application.

Here, we propose a simple model to estimate the planned water application (flooding duration) during Ag-MAR based on the following parameters: (1) soil texture; (2) crop saturation tolerance; (3) effective root-zone depth; and (4) critical water content. The concept of critical water (or air) content was proposed by several authors (Freijer, 1994; Glinski and Stepniewski, 1985; Hamamoto et al., 2011; Hunt, 2005; Moldrup et al., 2005; Troeh et al., 1982) as it indicates a percolation threshold where the gas transport path is blocked by pore-water, which results in gas diffusivity and permeability of practically zero. Hence, when the water content is either below or above this threshold, gaseous oxygen transport into the soil is blocked or opened, respectively (Fig. 1b). As opposed to the previous Ag-MAR models mentioned above, our proposed model is physically based and includes explicitly the soil water content, that is used to infer the soil aeration status. Yet, thanks to its simplicity, this model can be integrated easily into various existing Ag-MAR assessment tools such as SAGBI (O’Geen et al., 2015) or GRAT (https://www.groundwaterrecharge.org/).

In the following, we first describe the theory of the model and the methods used to test the model performance. Next, we present the model predictions and compare them with observations and numerical simulations. Last, we present an example of how to calculate Ag-MAR water application duration and we discuss the applicability of the model and its limitations.