Advances in the simulation of nutrient dynamics in cold climate agricultural basins: Developing new nitrogen and phosphorus modules for the Cold Regions Hydrological Modelling Platform
Introduction
Reducing nutrient losses from agricultural fields has been a major priority worldwide for many years due to increasing concerns with enhanced aquatic productivity and algal blooms. Water quality models for both basin and in-stream studies have been widely used to support nutrient management, but have often been problematic in seasonally cold regions such as Canada and the northern United States due to deficiencies in the representation of key processes specific to these regions. Cold regions hydrology cannot be represented by the classical concepts of rainfall-runoff models due to water storage by the seasonal snowcover, snow redistribution by wind, radiation-driven snowmelt, infiltration to and runoff over seasonally frozen ground, poorly defined drainage due to glacial geomorphology, and highly episodic runoff events (Pomeroy et al., 2007). Regional biogeochemistry in soils and runoff is challenging to model due to cold temperatures and seasonal soil freezing that influence nutrient release from soil–plant systems, plant uptake and microbial activity, which in combination with management practices (including fertilizer applications, tillage practices and wetland drainage) affect the hydrochemistry of soils and runoff (Baulch et al., 2019, Costa et al., 2020a, Irvine et al., 2019, Van Esbroeck et al., 2017, Macrae et al., 2007).
The dynamics of nutrient storage and release in cold climates are strongly affected by various cold regions hydrological processes and conditions (Deelstra et al., 2009). Snowpacks collect and transform chemicals during winter and rapidly release them during snowmelt (Pomeroy et al., 2005), with a significant portion of the nutrients contained in runoff being transformed and retained in topographical depressions (Neely and Baker, 1989, Crumpton and Isenhart, 1993, Birgand et al., 2007). Spring snowmelt is the largest runoff event of the year in cold regions such as the Northern Great Plains of North America (Gray et al., 1970), and accounts for most of the annual nutrient export (Baulch et al., 2019). The magnitude of peak flows furing spring freshet depends not only on overwinter snow accumulation but also on the antecedent soil moisture and basal snowpack and ground ice conditions (Gray et al., 1986, Pomeroy et al., 2007). Except for runoff from intensive convective rainfall events, summer flows are often small (Gray et al., 1970, Pomeroy et al., 2007).
Nitrogen (N) and phosphorus (P) transported via cold regions agricultural runoff originate in soil, vegetation, or to a lesser extent, the snowpack. The soil N pool is highly dynamic with weathering of soil parent material and decomposition of soil organic matter providing sources of mineral N (NO3-N and NH4-N) at rates depending on soil type and climate. Additional N enters the landscape through fertilizer application, plant residues, and atmospheric deposition. Transformations between labile and recalcitrant forms of N are generally biologically-driven with N lost to the atmosphere (through denitrification and volatilization) or to depth as soils drain (Baulch et al., 2011, Madramootoo et al., 2007). P exists in soils in both organic and inorganic forms, the latter derived from weathering of apatite. Like N, P enters the landscape through fertilizers, plant residues, and atmospheric deposition but is generally regarded to be less available due to soil sorption processes that are dependent on factors such as pH, temperature, and organic carbon content (Holtan et al., 1988). Phosphorus can be lost in runoff water, especially when concentrations exceed the sorption capacity of the soil, or when particulate P is transported along with soil through erosion processes. Soil frost can increase nutrient export by decreasing infiltration hence increasing soil–water interactions at the surface where soil P concentrations are often the highest (Cade-Menun et al., 2013). Additionally, freeze–thaw cycles disrupt plant cells and increase nutrient leaching from residues and other vegetation (White, 1973, Liu et al., 2013a, Costa et al., 2019a, Liu et al., 2019), which can become an important additional source of nutrients during snowmelt, particularly in the presence of young and actively growing plants (Cober et al., 2018, Elliott, 2013). The impact of tillage practices on nutrient export is complex. Conservation tillage can cause the accumulation of plant residue on farm fields, which can release nutrients to snowmelt runoff (Timmons et al., 1970, Miller et al., 1994, Ulén, 1997). In addition, by decreasing the mixing of the applied fertilizer, reduced tillage increases nutrient soil stratification and can lead to higher nutrient concentrations in surficial soils, which can be readily mobilized by runoff.
More reliable predictions of nutrient transport in cold agricultural basins have long been seen as a crucial to support nutrient management in Canada (Costa et al., 2020a, Baulch et al., 2019, Costa et al., 2019b). (Mekonnen, 2016) identified 74 models of water quality worldwide, but it has been noted that application of many of these models can be problematic in cold climates due to inadequate representation of many cold regions processes (Han et al., 2010). Costa et al. (2020a) reviewed the suitability of five prominent catchment nutrient models for application in cold climates: SWAT (Arnold et al., 1998), INCA (Whitehead et al., 1998, Wade et al., 2002), HYPE (Lindström et al., 2010, Arheimer et al., 2012), HSPF (Bicknell et al., 1997, Bicknell et al., 2005, Duda et al., 2012), and AnnAGNPS (Bosch et al., 1998). They identified inadequate representation of cold climate hydrology and daily time steps to be some of the features most commonly limiting the utility of these models in cold regions. They noted that most of these models have rarely been applied to cold regions, with the exception of SWAT and HYPE. They also found that some models allowed limited soil vertical resolution (i.e., maximum number of soil layers) that could reduce their performance in heavily stratified soils. Erosion remains a major challenge and meaningful model structures based on observable and transferable parameters were recommended to reduce the often high number of parameters for controlling biogeochemical transformations (leading to parameter identifiability). It was also highlighted that representations of accumulation of immobile nitrogen and phosphorus organic pools were often limited in their ability to represent legacy N and P for long-term simulations.
The meteorological data typically used to force hydrological models (e.g. solar radiation, air temperature, precipitation and wind speed) are often measured on a daily basis. This may limit the temporal resolution of model simulations and compromise their ability to capture hydrological and transport-biogeochemical processes that may be subject to significant diurnal variations (e.g. wind redistribution of snow and radiation-driven snowmelt) and episodic oscillations (e.g., sediment erosion and soil nutrient release). Unfortunately, in cold regions, hydro-biogeochemical processes activated during spring snowmelt and convective storms are often responsible for most of the annual nutrient export (Baulch et al., 2019, Kokulan et al., 2019). Thus, long-term simulations must also capture short-term runoff events, meaning that daily timestep models are insufficient. For example, the HSPF model is one of the few nutrient models identified that are often run at (default) hourly time intervals for long-term simulations. However, like many other models, HSPF does not explicitly account for some critical cold regions processes such as blowing snow and infiltration to and runoff over frozen soils, and uses the daily time step empirical degree-day method to estimate snowmelt, a method that has long been found to be inadequate in many cold regions (Walter et al., 2005, Gray and Landine, 1988).
There is a need to investigate alternative modelling approaches that are more applicable to cold agricultural basins, better reflecting cold regions hydrological and biogeochemical processes, which have a crucial impact on timing, concentration and load of nutrients. For this purpose, a complex of new hydro-biogeochemical modules was developed for the flexible, modular Cold Regions Hydrological Modelling platform (CRHM). CRHM has been created specifically to improve the simulation of cold regions hydrology (Pomeroy et al., 2007) and has been applied successfully to agricultural basins with minimal or no calibration (Fang and Pomeroy, 2008, Fang et al., 2010, Mahmood et al., 2017, Cordeiro et al., 2017, Costa et al., 2017, Kokulan et al., 2019). Its merit as a flexible and fundamentally physically based cold regions hydrological model renders it an ideal model for incorporating nutrient processes, and hence to offer a more suitable modelling framework to support nutrient management in agricultural cold regions.
Section snippets
The Cold Regions Hydrological Model
The Cold Regions Hydrological Model (CRHM) has been developed from more than 55 years of research on Canadian hydrology (Pomeroy et al., 2007). It is a modular platform that discretizes the basin into hydrologically distinct landscape elements called Hydrological Response Units (HRUs). It provides a range of predictive methods embedded in various modules that can be selected depending on dominant climatic and regional settings, i.e., mountains and prairie environments and are applied to
Hydrology
Fig. 4 compares observed and simulated SWE in fields F3 and F4 (locations are shown in Fig. 3). The observations correspond to the average snow accumulation peaks measured at the onset of spring snowmelt. The results show that the model can capture both the interannual and spatial variabilities in SWE distribution. Substantial heterogeneity in annual snow accumulation can be noticed within each field - note the standard deviation (error bars) for each year as a measure of the spatial variation
Critical management of timing of fertilizer use relative to major runoff events
Spring snowmelt is frequently the major annual nutrient export event in the Canadian Prairies, but fertilizer and manure applications in the growing season can also be mobilized via summer and spring rainfall-runoff events (Nicholaichuk, 1967, Hansen et al., 2002, Glozier et al., 2006, Liu et al., 2013b). This model and model application show that the timing of nutrient applications plays a key role in the amount of nutrients exported via runoff in southern Manitoba. Fig. 10 highlights that by
Conclusions
A series of process-based transport and biogeochemical modules have been developed for the Cold Regions Hydrological Model (CRHM) to simulate nitrogen (N) and phosphorus (P) in cold agricultural basins. The new model aims to address critical issues with existing nutrient models for simulation of these environments.
The new modules calculate nutrient fluxes throughout the basins’ hydrological system that includes the snowpack, soil, streams and depressional storage (e.g., potholes and wetlands).
Data Availability Statement
The data that support the findings of this study are available from Agriculture and Agri-Food Canada (AAFC) and Environment and Climate Change Canada (ECCC). Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the authors with permission of AAFC and ECCC.
CRediT authorship contribution statement
Diogo Costa: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. John W. Pomeroy: Conceptualization, Formal analysis, Funding acquisition, Supervision, Software, Writing – review & editing. Tom Brown: Formal analysis, Software. Helen Baulch: Supervision, Writing – review & editing. Jane Elliott: Supervision, Writing – review &
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.
Acknowledgments
This research was supported by the Global Water Futures Program and Environmental and Climate Change Canada. The authors would like to thank Agriculture and Agri-Food Canada and Environment and Climate Change Canada for kindly providing the data used in this study. The research would not have been possible without the interest and cooperation of landowners in the South Tobacco Creek Basin and the Deerwood Soil and Water Management Association.
References (100)
- et al.
Wetland drainage in the Canadian prairies: Nutrient, salt and bacteria characteristics
Agriculture, Ecosystems & Environment
(2012) - et al.
A numerical model for the simulation of snowpack solute dynamics to capture runoff ionic pulses during snowmelt: The PULSE model
Advances in Water Resources
(2018) - et al.
Evaporation from natural nonsaturated surfaces
Journal of Hydrology
(1989) - et al.
Seasonal nutrient export dynamics in a mixed land use subwatershed of the Grand River, Ontario, Canada
Journal of Great Lakes Research
(2019) The seasonal snowcover
- et al.
Modeling of nutrient export and effects of management practices in a cold-climate prairie watershed: Assiniboine River watershed, Canada
Agricultural Water Management
(2017) - et al.
The effect of sampling frequency and strategy on water quality modelling driven by high-frequency monitoring data in a boreal catchment
Journal of Hydrology
(2019) - et al.
The rapid eutrophication of Lake Winnipeg: Greening under global change
Journal of Great Lakes Research
(2012) - et al.
The transformation of frequency distributions of winter precipitation to spring streamflow probabilities in cold regions; case studies from the Canadian Prairies
Journal of Hydrology
(2015) Nutrient losses by surface run-off from soils with winter cover crops and spring-ploughed soils in the south of Sweden
Soil and Tillage Research
(1997)