Integrating statistical and hydrological models to identify implementation sites for agricultural conservation practices☆
Introduction
Maintaining sustainability of water resources while fulfilling the demand in global food and energy security is a challenging task, especially in light of increased global population and extreme climatic events (Jordan et al., 2012, Giri, 2013). Specifically, maintaining acceptable water quality is necessary to protect human, animal, and plant health. Preventing water quality degradation due to non-point source (NPS) pollution is difficult because of its diffuse and complex nature (Ouyang et al., 2009, Ding et al., 2010, Laia et al., 2011). According to the United States Environmental Protection Agency (USEPA), agriculture is the primary contributor of NPS pollution to rivers and lakes (USEPA, 2005). Agricultural NPS pollutants include sediment, nutrients, pathogens, organic materials, toxic chemicals and pesticides (USEPA, 1992). Excessive sediment loading in water bodies reduces dissolved oxygen levels and increases water temperature, negatively impacting aquatic organisms (Malone, 2009). An increase in nutrients such as nitrogen and phosphorus accelerates eutrophication, which is harmful to aquatic organisms. In addition, toxic chemicals and pesticides cause long-term contamination of surface and subsurface water resources and threaten both human and aquatic ecosystems (Love et al., 2011).
Best management practices (BMPs) are well known methods to reduce pollutants in surface runoff generated from agricultural lands (Arabi et al., 2007, Sommerlot et al., 2013a). The pollution reduction mechanism in BMPs follows three principles: 1) decreasing pollutant concentration through nutrient and pesticide management while reducing pollutant mass by erosion control practices; 2) reducing pollutant transport into waterbodies through various vegetative barriers; and 3) lowering pollutant mass through chemical and biochemical removal processes (Cunningham et al., 2003). However, the efficiency of BMPs varies within a watershed because it depends on various factors, including topography, soil characteristics, geological formations, climate, cropping systems, and cultural practices.
Efficiency of BMPs can be evaluated using monitoring or modeling, with monitoring being more accurate than modeling. However, monitoring water quality data throughout a watershed is time consuming, expensive, and often ineffective. In order to address these difficulties, modeling that utilizes currently available data is preferred because it is faster and less expensive than monitoring. Several studies have used physically-based watershed models to evaluate BMP effectiveness in improving water quality (Wang et al., 2009, Panagopoulos et al., 2012, Ciou et al., 2012, Amon-Armah et al., 2013, Lescot et al., 2013, Liu et al., 2013, Qiu, 2013, Sommerlot et al., 2013b). However, these models cannot be used effectively unless created with stakeholders' knowledge and used in development of a watershed management plan (Nejadhashemi et al., 2011, Sun, 2013). Involvement of stakeholders in each step of the watershed management plan is essential, as their social, economic, and environmental knowledge helps in achieving any proposed solution (Nejadhashemi et al., 2009). In addition, the likelihood of a successful watershed management plan is much higher if more effective stakeholder involvement takes place (Bosch et al., 2012). Effective stakeholder involvement occurs when they can employ scientific tools and data related to watershed processes and management (Maguire, 2003). However, most physically-based watershed models can be only used effectively by experts who have adequate training in agricultural science and significant knowledge in computer modeling (Saleh et al., 2011), making them less attractive to stakeholder groups.
In this study, the overall goal is to facilitate information transfer to stakeholders by introducing simpler techniques that predict pollution reduction at the watershed outlet using watershed characteristics. This would replace the use of complex models by watershed managers and stakeholders in the planning and decision making process, which can be used in agricultural watersheds around the world. We hypothesized that BMP effectiveness varies by BMP implementation site, which can be described either by its distance to the watershed outlet or the stream order concept. The specific objectives of the study were to 1) develop an empirical model to predict BMPs' pollution reduction using watershed characteristics such as distance to the watershed outlet and stream order, and 2) visualize the overall watershed response to pollution reduction and determine the optimal distance of BMP implementation from the watershed outlet.
Section snippets
Study area
This study was conducted in the Saginaw River Watershed located in east-central Michigan with a total area of 15,263 km2 (Fig. 1). The watershed drains northeast into Lake Huron. This watershed contains the largest contiguous freshwater coastal wetland of the nation and is one of the most diverse watersheds in Michigan (USEPA, 2009). The Saginaw River Watershed consists of 23% agricultural land, primarily corn and soybean. The remaining is 42% forest, 17% pasture, 11% wetlands, and 7% urban
Data distribution and correlation
Fig. 2 shows the Spearman's correlation matrix. In this figure the scatterplots are located in the lower panel, Spearman's rank correlation coefficients are presented in the upper panel, and histograms are presented on the diagonal. The red line in the scatterplots depicts the simple linear regression between a variable pair. The histograms represent the observed distribution of the data, while Spearman's rank correlation coefficients represent the correlation between two variables. All
Conclusion
Although physically-based models such as SWAT are accurate and reliable tools for watershed management, require expert knowledge of watershed processes to be used properly. Therefore, these types of models are not widely used beyond a planning stage and cannot be adjusted by stakeholders to examine new management scenarios without the help of experts. Our goal in this study was to develop user-friendly models that can facilitate information transfer of advanced models to non-expert users.
The
Acknowledgments
This work is supported by the USDA National Institute of Food and Agriculture, Hatch project MICL02212.
References (44)
- et al.
Community decision: stakeholder focused watershed planning
J. Environ. Manag.
(2012) - et al.
Development and test of the export coefficient model in the upper reach of the Yangtze River
J. Hydrol.
(2010) - et al.
Evaluation of targeting methods for implementation of best management practices in the Saginaw River Watershed
J. Environ. Manag.
(2012) - et al.
Catchment science and policy for agriculture and water quality
Environ. Sci. Policy
(2012) - et al.
A spatially distributed cost effectiveness analysis framework for controlling water pollution
Environ. Modell. Softw.
(2013) - et al.
Assessment of effects of best management practices on agricultural non-point source pollution in Xiangxi River watershed
Agric. Water Manag.
(2013) - et al.
Effects on aquatic and human health due to large scale bioenergy crop expansion
Sci. Total Environ.
(2011) - et al.
Scale-dependent soil and climate variability effects on watershed water balance of the SWAT model
J. Hydrol.
(2002) - et al.
Temporal-spatial dynamics of vegetation variation on non-point source nutrient pollution
Ecol. Modell.
(2009) - et al.
Decision support for diffuse pollution management
Environ. Modell. Softw.
(2012)
Evaluating the capabilities of watershed scale models in estimating sediment yield at field scale
J. Environ. Manag.
Evaluating the impact of field-scale management strategies on sediment transport to the watershed outlet
J. Environ. Manag.
Enabling collaborative decision-making in watershed management using cloud-computing services
Environ. Modell. Softw.
A global sensitivity analysis tool for the parameters of multi-variable catchment models
J. Hydrol.
A new look at the statistical model identification
IEEE Trans. Autom. Control.
Effect of nutrient management planning on crop yield, nitrate leaching and sediment loading in Thomas Brook Watershed
Environ. Manag.
Representation of agricultural conservation practices with SWAT
Hydrol. Process.
Large area hydrologic model development and assessment part 1: model development
J. Am. Water Resour. Assoc.
Package Language R
AIC model selection and multimodal inference in behavioral ecology: some background, observations, and comparisons
Behav. Ecol. Sociobiol.
Optimization model for BMP placement in a reservoir watershed
J. Irrig. Drain. Eng. ASCE
An Assessment of the Quality of the Agricultural Best Management Practices Implemented in the James River Basin of Virginia
Cited by (16)
Climate change vulnerability assessment and adaptation strategies through best management practices
2020, Journal of HydrologyCan soil conservation practices reshape the relationship between sediment yield and slope gradient?
2020, Ecological EngineeringAssessing the potential impacts of climate and land use change on water fluxes and sediment transport in a loosely coupled system
2019, Journal of HydrologyCitation Excerpt :Better understanding of how the changes in land use and climate will impact the hydrological cycle, and especially sediment transport, will enable more informed watershed adaptation strategies to help make communities more resilient to climate change. To that end, model simulation is proven to be an effective tool (Woznicki et al., 2016; Giri et al., 2015). A variety of models have been developed to assess the impact of climate and land use change on watershed scale, however, Soil and Water Assessment Tool (SWAT) has seen widespread application due to its process based structure, ability to model hydrology, plant growth related process, incorporate different urban and agricultural management practices, representation of land use and meteorological parameters essential from water balance prospective (Zhang et al., 2018; Giri et al., 2016a; Ficklin and Barnhart, 2014; Giri et al., 2014; Mutenyo et al., 2013).
Evaluation of the effectiveness of conservation practices under implementation site uncertainty
2018, Journal of Environmental ManagementEvaluating the role of evapotranspiration remote sensing data in improving hydrological modeling predictability
2018, Journal of HydrologyCitation Excerpt :A widely used stochastic weather generator called WXGEN was employed (Sharpley and Williams, 1990; Wallis and Griffiths, 1995), which is embedded in the Soil Water Assessment Tool (SWAT), to create climate time series for other climatological records (e.g. relative humidity, solar radiation, and wind speed) that are required for SWAT to operate (Neitsch et al., 2011). Predefined crop management operations, schedules, and rotations were adopted from previous studies performed in the same region (Love and Nejadhashemi, 2011; Giri et al., 2015). Due to limitation of SWAT in simulating up to 250 different landuse, the subwatershed map that was provided by the National Hydrology Dataset plus (NHDPlus) and the Michigan Institute for Fisheries Research at a scale of 1:24,000 were modified to accommodate this limitation (Einheuser et al., 2013).
- ☆
Special Issue on Agricultural Systems Modeling & Software.