A case study on the effects of data temporal resolution on the simulation of water flux extremes using a process-based model at the grassland field scale
Introduction
Flooding in the UK puts more than 5 million people in 2.4 million properties at risk each year (Environment Agency, 2009). Projected changes to rainfall patterns (Watts and Anderson, 2016) may exacerbate the existing risks posed by flooding. Flash flooding or surface water flooding, defined as those flood events where the rise in water is either during or within a few hours of the rainfall that produces the rise, is one of the most common types of flooding in the UK. The utilised agricultural area, of which almost 60% is permanent grassland, covers 71% of the total land of the UK (Department for Environment, Food and Rural Affairs, 2019). Water fluxes or surface runoff generated from agricultural land can contribute significantly to local floods and nutrient losses that cause water pollution. Flooding of farmland is likely to become more frequent in some areas under projected climate change (Brown et al., 2016), although intriguingly, studies have found that increases in precipitation extremes do not necessarily mean increases in flood magnitude, due to decreased soil moisture at storm onset and reduced storm durations (Sharma et al., 2018, Wasko et al., 2019). Further, soil erosion is accelerating due to more intense rainfall, leading to the loss of valuable topsoil and the pollution of watercourses (Morison and Matthews, 2016).
Accurate forecasting of water runoff (or water fluxes) from agricultural land is, therefore, not only a vital component of flood early-warning systems, but also important for associated management strategies for nutrient loss and water pollution. Water fluxes from the soil surface are controlled by soil properties. Long-term hydrological studies have shown that sandy Alfisols can generate higher runoff compared to clayey Vertisols (Pathak et al., 2013), and a greater risk of flooding on clay soils has been reported (Charlton et al., 2010). The wetness of the soil before a precipitation event (Merz and Plate, 1997) and soil compaction also affect water fluxes. Farm machinery and livestock (Adimassu et al., 2019, Alaoui et al., 2018, Newell Price et al., 2012) can cause serious compaction and so exacerbate flood risk, and natural events, particularly long and intense precipitation events (Archer and Fowler, 2018), and land cover variation (Dadson et al., 2017, Keesstra et al., 2018) also affect flux.
Agricultural systems are complex because they are generally managed at the field scale and each field has its own unique set of soil conditions and topology. Monitoring water surface fluxes in fields is costly both in time and financially. In this respect, modelling provides an effective tool for simulating or forecasting water fluxes. The SPACSYS model (Wu et al., 2007) is one such process-based model. It is a field scale and weather-driven dynamic simulation model. Since it was first published in 2007, it has been developed to provide added functionality (Bingham and Wu, 2011, Liu et al., 2013, Wu et al., 2019, Wu et al., 2015). The model can simulate the interactions of soil carbon (C), nitrogen (N) and phosphorus (P), plant growth and development, water re-distribution and heat transformation in agricultural fields. The model has been used to investigate several issues including resource use efficiency by crops (Wu et al., 2009), greenhouse gas (GHG) emissions (Abalos et al., 2016, Perego et al., 2016), the responses of cropping and grassland systems to environmental change (Wu et al., 2016), and the forecasting of crop yield and stocks of C and nutrients (Zhang et al., 2016) under various climatic and soil conditions.
The SPACSYS model has been developed to investigate not only temporal dynamics, but also within-field spatial variation in processes such as water runoff, using a linked, grid-based approach (grid-to-grid) (Liu et al., 2018). As in all previous implementations of SPACSYS, and common to many agriculture-focused models (Ahuja et al., 2002), a daily time-step was used. However, model predictions of water flux did not increase in accuracy when considering grid connectivity. We hypothesise, that a finer time-step might provide this improvement instead; not only in the grid-to-grid model, but also in the (non-grid-to-grid) standard model, as investigated here. Although not demonstrated within this study, increasing the accuracy of water flux simulations should implicitly increase the accuracy of associated SPACSYS simulations, such as those for nutrient loss that use predicted water flux in their calculation.
For our case study, we used measured 15-minute water flux data from one field (or sub-catchment) of the North Wyke Farm Platform (NWFP). The NWFP is a systems scale research facility in the south-west of England for investigation of the sustainability of lowland ruminant production systems (Orr et al., 2016). South-west England has a relatively wet climate where the greatest rainfall is in winter and the driest times are between April to July. August tends to show an increase in rainfall over July and starts the inexorable rise in rainfall into autumn and early winter. More recently, the number of flood events has increased (Stevens et al., 2016), mostly in the autumn and winter months; all as a likely consequence of increased surface water runoff (Palmer and Smith, 2013).
For this study, the NWFP’s 15-minute water flux data were up-scaled to hourly, 6-hourly and daily data and the SPACSYS model was adapted to provide corresponding downscaled simulations at 15-minute, hourly and 6-hourly resolutions (in addition to its usual daily output). This provided four measured water flux datasets and four simulated water flux datasets over a study period of 34 months (April 2013 to February 2016). Simulations were generated using the same field management practices and parameter configurations. These rich water flux datasets enabled investigation of the effects of temporal scale on model performance not only in terms of extreme water runoff, which is the focus of this study and provides its novelty, but also in terms of general trends.
Section snippets
Model description
The SPACSYS model includes a plant growth and development component, an N cycling component, a C cycling component, a P cycling component, plus a soil water component that includes representation of water flow to field drains as well as downwards through the soil layers, together with a heat transfer component. The equations to quantify such different processes have been described elsewhere (Liu et al., 2013, Wu et al., 2019, Wu et al., 2007, Wu et al., 2015). Here, only the processes
Model performance for each of the four data temporal resolutions, separately
Comparisons between the measured and simulated water flux rates at different temporal resolutions are shown in Fig. 2. Visually, it appears that simulations of daily and 6-hourly water fluxes tend to under-predict the measured data, often missing high peaks, while simulations of 15-minute and hourly data possibly tend to over-predict. However, the scatterplots of the measured and simulated data, together with the ideal 1:1 line, a linear regression fit, and a Loess smoother fit (Fig. 3) present
Unaggregated data
The statistical analyses for model performance suggested that the SPACSYS model simulates the general trend of water fluxes at the four different temporal resolutions reasonably well (Fig. 2, Fig. 3). All simulations tended to over-predict water flux, however, and only simulations at the finest resolutions maintained the variation in the measured data. The accuracy of water flux peak simulations varied among the four resolutions (Table 2). Almost 92% of the measured peaks over the simulated
Conclusions
For the grassland study site, the adapted process-based model (SPACSYS) was able to adequately simulate the trends in measured water fluxes and identify their extremes. At a daily time-step, model accuracy increased when simulations were run at finer temporal resolutions, specifically 15-minute and hourly, and then aggregated to daily (a coarse output resolution commonly used in field-scale agricultural settings). Aggregating using 6-hourly simulations was less accurate. For the study site,
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This research was funded by the BBSRC Institute Strategic Programme grant, “Soils to Nutrition” (BBS/E/C/000I0330, BBS/E/C/000I0320), the BBSRC National Capability grant for the North Wyke Farm Platform (BBS/E/C/000J0100) and a PhD studentship funded by Rothamsted Research and Lancaster University.
References (57)
- et al.
Climate change and N2O emissions from South West England grasslands: a modelling approach
Atmos. Environ.
(2016) - et al.
Effects of tillage and crop residue management on runoff, soil loss and crop yield in the Humid Highlands of Ethiopia
Agric. Syst.
(2019) - et al.
Does soil compaction increase floods? A review
J. Hydrol.
(2018) - et al.
Simulation of wheat growth using the 3D root architecture model SPACSYS: validation and sensitivity analysis
Eur. J. Agron.
(2011) - et al.
Impact of transition from permanent pasture to new swards on the nitrogen use efficiency, nitrogen and carbon budgets of beef and sheep production
Agric. Ecosyst. Environ.
(2019) - et al.
An evaluation of automated GPD threshold selection methods for hydrological extremes across different scales
J. Hydrol.
(2020) - et al.
Global scale DAYCENT model analysis of greenhouse gas emissions and mitigation strategies for cropped soils
Glob. Planet. Change
(2009) - et al.
Impact of temporal resolution of inputs on hydrological model performance: an analysis based on 2400 flood events
J. Hydrol.
(2016) - et al.
The superior effect of nature based solutions in land management for enhancing ecosystem services
Sci. Total Environ.
(2018) - et al.
Modelling field scale spatial variation in water run-off, soil moisture, N2O emissions and herbage biomass of a grazed pasture using the SPACSYS model
Geoderma
(2018)
Hydrological behavior of Alfisols and Vertisols in the semi-arid zone: Implications for soil and water management
Agric. Water Manag.
Field evaluation combined with modelling analysis to study fertilizer and tillage as factors affecting N2O emissions: a case study in Po valley (Northern Italy)
Agric. Ecosyst. Environ.
Field-based determination of controls on runoff and fine sediment generation from lowland grazing livestock fields
J. Environ. Manag.
Detecting land use and land management influences on catchment hydrology by modelling and wavelets
J. Hydrol.
The role of spatial and temporal model resolution in a flood event storyline approach in western Norway
Weather Clim. Extrem.
Sensitivity of Australian roof drainage structures to design rainfall variability and climatic change
Build. Environ.
SPACSYS: integration of a 3D root architecture component to carbon, nitrogen and water cycling - model description
Ecol. Model.
Simulation of nitrous oxide emissions at field scale using the SPACSYS model
Sci. Total Environ.
Effects of enhancing soil organic carbon sequestration in the topsoil by fertilization on crop productivity and stability: evidence from long-term experiments with wheat-maize cropping systems in China
Sci. Total Environ.
Agricultural System Models in Field Research and Technology Transfer
Characterising flash flood response to intense rainfall and impacts using historical information and gauged data in Britain
J. Flood Risk Manag.
Preliminary study of computational time steps in a physically based distributed rainfall–runoff model
Water
Robust locally weighted regression and smoothing scatterplots
J. Am. Stat. Assoc.
UK Climate Change Risk Assessment 2017
A restatement of the natural science evidence concerning catchment-based 'natural' flood management in the UK
Proc. R. Soc. A
Cited by (3)
Impact of plastic mulching and residue return on maize yield and soil organic carbon storage in irrigated dryland areas under climate change
2024, Agriculture, Ecosystems and EnvironmentExploring the effects of land management change on productivity, carbon and nutrient balance: Application of an Ensemble Modelling Approach to the upper River Taw observatory, UK
2022, Science of the Total EnvironmentCitation Excerpt :More details about the model are presented in S.I. The SPACSYS model can run on a range of time steps but uses a daily time-step in this study (Wu et al., 2021). Parameters have been calibrated and validated near the study site as well as other sites with different climate and soil conditions for grazing livestock (e.g. Carswell et al., 2019; Wu et al., 2016), silage (e.g. Sándor et al., 2020) and cereal and legume crops (e.g. Bingham and Wu, 2011; Liang et al., 2019; Liu et al., 2013).