Land Management Strategies for Watershed Restoration: An Integration of Spatial Modeling with Dynamic Programming

Final Report for GNC07-073

Project Type: Graduate Student
Funds awarded in 2007: $9,004.00
Projected End Date: 12/31/2008
Grant Recipient: The Ohio State University
Region: North Central
State: Ohio
Faculty Advisor:
Brent Sohngen
The Ohio State University
Expand All

Project Information

Summary:

Conservation management practices are considered one of the best answers to escalating water quality deterioration by nonpoint source pollution. Integrated watershed economic model (IWEM) offers a multidisciplinary framework by addressing both the biophysical and the economic (cost and benefit) aspects of water quality improvement. An IWEM can be conceptualized as three sub-models: a watershed model, an economic model, and an optimization tool to integrate the watershed and economic models together. The present study is an attempt in this direction, by translating the three sub-models of IWEM into three essays of the dissertation. The Upper Big Walnut Creek (UBWC) watershed in central Ohio was selected for applying the IWEM framework. The modeling of the UBWC watershed was performed in the first essay. For this study the Soil and Water Assessment Tool (SWAT) was used to predict the water quality changes associated with land management practices. A dynamic programming-based economic optimization approach was used in this study, which could capture the nutrient movements in agro-ecosystems, starting from nutrient application, intake by plants and transport from the field to downstream water reservoir with possible nutrient assimilation in-between. The social cost of the pollution is parameterized with benefit estimates of water quality improvement. Model is developed for the entire watershed by considering it as a single homogeneous one hectare unit. The watershed model was used to simulate the baseline, and crop rotation and conservation technology-specific production functions. Two sets of conservation technologies were developed for the watershed. One with split nitrogen fertilizer application, cover cropping, conservation tillage and vegetative buffer stripes and the other with 25% reduction in nitrogen fertilizer, cover cropping, conservation tillage and vegetative buffer stripes. The analysis revealed that under no restriction on N loading, farmers would apply a maximum of 170.51kg/ha of N and the value function would be $7950 under C-S-W rotation. However, after introducing the social cost of pollution in objective function, the fertilizer application rate was reduced to 103 kg/ha. Additionally, within the crop-technology combination, technology Set-3(split-N application, conservation tillage, cover crop and vegetative buffer) showed the lowest pollution load to the reservoir along with higher value function.

Introduction:

Today, non-point source pollution (NPS) is one of the major sources of water quality impairments globally (UNEP, 2007). In the US, nutrient pollution is the leading cause of water quality issues in lakes and estuaries (USEPA, 2002). The maximum concentration of nutrients in streams is found to be in agricultural basins, and it is correlated with nutrient inputs from fertilizers and manures. This clearly shows the role of agricultural practices in water quality degradation (USGS, 1999). To improve the quality of water bodies, the United States Environmental Protection Agency (USEPA) mandates individual states to implement the Total Maximum Daily Load (TDML) (USEPA, 2002). The state and federal governments are working with several conservation programs to reduce the NPS load from agriculture (Mausbach and Dedrick, 2004). However, the ever-increasing water quality impairment by agricultural NPS in US clearly shows that the task of formulating and implementing the cost-effective policies for controlling the NPS impact on water resources is challenging.

Project Objectives:

The specific objectives of the projects are to,

1. to Identify, evaluate and prioritize the site-specific on-farm and off-farm land management strategies to reduce nitrogen loading using SWAT model.

2. to optimize the land management strategies to maximize social benefit

Cooperators

Click linked name(s) to expand/collapse or show everyone's info
  • Brent Sohngen

Research

Materials and methods:

An integrated watershed-economic modeling (IWEM) offers a holistic approach, where compounding effect of biophysical and anthropogenic variables can be identified and their impact on NPS can be partitioned by linking the biophysical process and the economic behavior models. Such an IWEM would have three components, a biophysical process model component, an economic behavior component and a tool to integrate both the biophysical and economic components. In this research IWEM methodology is applied to the Upper Big Walnut Creek (UBWC) watershed of central Ohio to derive socially benefiting choices of conservation practices to reduce nutrient nitrogen (N) load from agriculture. The UBWC watershed was identified by Ohio EPA as an impaired watershed due to nutrient enrichment from agricultural (Ohio EPA, 2005). Additionally, the watershed encompasses perennial and intermittent streams that drain into Hoover Reservoir, and serves as a primary source of drinking water supply and a favorite local recreational site for residents in the neighboring communities.

Soil and Water Assessment Tool (SWAT), a widely used basin scale biophysical process model was used as biophysical component of IWEM. The baseline nutrient production function, watershed level N production function for corn and wheat, and phosphorous production function for soybean were estimated by using SWAT model for the UBWC watershed. A quadratic relationship between applied nutrients and the yield were established by regressing applied nutrient against simulated yields of for different crops for the watershed. In addition, SWAT model was also used to derive the baseline soil N balance equation. The conservation management options, such as split application of N fertilizer, conservation tillage, cover cropping and vegetative buffer were simulated using the SWAT model for deriving crop and technology specific quadratic nutrient production functions and N loading function. The predominant crop rotations in the watershed, corn-corn (C-C), corn-soybean (C-S) and corn- soybean-wheat (C-S-W) were considered for SWAT simulations.

The economic component of IWEM consists of social cost of N load, cost of production of crops and technology cost of conservation practices. The benefits of water quality improvements were derived from two different studies, 1) The recreational value of water quality improvement were estimated based from a combined stated and revealed preference method applied to UBWC and 2) A conjoint analysis of the use (excluding recreation) and non-use value of water quality improvement in UBWC reported by Tennity (2005). The value of complete marginal benefit of per hectare N loading reduction by half from a farm was estimated as $328.77 for streams and $387.86 for Hoover reservoir, which was used to parameterize social damage cost (SDC) of N loading for UBWC. SDC was assumed as a power function of N loading and the elasticity parameter was fixed as 2 after series of simulation by different values for elasticity and intercept parameter for SDC. The benefit estimates of N load reduction and elasticity parameter value of 2 were used to fix the intercept parameter of SDC, as 0.101 for in-stream and 0.19 for SDC at Hoover reservoir in UBWC. Additional cost involved in adoption of conservation technologies were obtained from different source (Hoorman, 2009; Sohngen 2003).

As different conservation practices were applied simultaneously by a farmer. Three different technology sets were considered for DP based scenario analysis 1) Technology Set-1: The current level of agricultural production and N loading 2) Technology set-2: Cover crop, vegetative buffer and conservation tillage and 3) Technology set-3: Technology set-2 + split-N fertilizer application. The changes in crop production and N-loading for technology set-2 and 3 were expressed as an exponential function of level of adoption of from the baseline (Technology Set-1) crop production and N-loading. Three different DP problems were specified for C-C, C-S and C-S-W crop rotations. In the case of C-S rotation and C-S-W rotations, total profit from each crops were weighted with the proportion of area under each crop. The DP problem was specified as deterministic with finite horizon and two state variables, N stock in the soil and N stock in the downstream reservoir. Fertilizer application was considered as action variable of DP (assuming 100% technology adoption). Thus, one action variable (N application for corn) for C-C rotation, two action variable for C-S rotation, N fertilizer application for corn and P fertilizer application for soybean and three action variable for C-S-W rotation, N fertilizer application for corn and wheat and P fertilizer application for soybean. The Bellman equation expressed as an internalized profit function for N loading. Each of the dynamic programs was sequentially run for different technology scenarios.

Research results and discussion:

The DP analysis for baseline crop production results were close to the Ohio field crop enterprise budget. In addition, N loading in baseline simulations was also in line with the modeled of the watershed. The analysis revealed that under no restriction on N loading, farmers would apply a maximum of 170.51kg/ha of N and the value function would be $7950 under C-S-W rotation. However, after introducing the social cost of pollution in objective function, the fertilizer application rate was reduced to 103 kg/ha. The analysis of conservation management options revealed that each of the crop rotation and technology combination would give higher value than the present level of production with internalized pollution cost. Within the crop-technology combination, technology Set-3(split-N application, conservation tillage, cover crop and vegetative buffer) showed the lowest pollution load to the reservoir along with higher value function. From the results it could conclude that the present level of private profit and yield levels are not realized by adopting both the technology sets considered in this study. Additionally, more area under C-C and C-S rotation would result in more pollution load to the reservoir.

Participation Summary

Educational & Outreach Activities

Participation Summary:

Education/outreach description:

Surendran Nair. S, K.W. King, J.D. Witter, B.L. Sohngen, and N.R. Fausey. Extending beyond discharge in calibrating watershed water quality simulation tools. (Under review in Journal of American Water Resource Association).

Surendran Nair, S., King, K., Witter, J., Sohngen , B., Fausey, N. 2011. Importance of crop yield in calibrating watershed water quality simulation tools. Accepted for oral presentation at Tennessee Water Resources Symposium, April 2011.

Surendran Nair. S., B. Sohngen, K. King, N. Fausey, J. Witter. 2010. Integrated Watershed Economic Model for Non-Point Source Pollution Management in the Upper Big Walnut Watershed. Oral Presentation at the AAEA Meeting, Denver, CO. July, 2010.

Surendran Nair. S., B. Sohngen, N. Fausey, K. King, 2007.Optimal Management of Non-point Source Pollution from Agriculture: An Application of Dynamic Programming. Oral Presentation at the SWCS Conference, Tampa, FL. July, 2007.

Project Outcomes

Project outcomes:

The analysis of conservation management options revealed that technology Set-3(split-N application, conservation tillage, cover crop and vegetative buffer) showed the lowest pollution load to the reservoir along with higher value function. From the results it could conclude that the present level of private profit and yield levels are not realized by adopting both the technology sets considered in this study. Additionally, more area under C-C and C-S rotation would result in more pollution load to the reservoir.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.