Two reactionary approaches to reduce impaired agricultural sediment and nutrient runoff

Open Access
- Author:
- Jiang, Fei
- Graduate Program:
- Soil Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 25, 2020
- Committee Members:
- Patrick Joseph Drohan, Dissertation Advisor/Co-Advisor
Patrick Joseph Drohan, Committee Chair/Co-Chair
Cibin Raj, Outside Member
Charles White, Committee Member
Scott A Roths, Outside Member
Heather Elise Preisendanz, Committee Chair/Co-Chair
David Eissenstat, Program Head/Chair
Heather Elise Preisendanz, Dissertation Advisor/Co-Advisor - Keywords:
- riparian buffer
ecosystem services
agricultural water quality
land use optimization
modeling
SWAT
REMM
best management practices
crop yield
nutrient and sediment loss
land suitability - Abstract:
- Excessive nitrogen, phosphorus and sediment loads from agricultural runoff continue to hamper stream water quality despite decades of land management efforts. Current legislative mandates, while a promising step forward, are also having limited effect. For example, Pennsylvania (PA), US and Northern Ireland (NI) face challenges in meeting water quality standards as mandated by the Chesapeake Bay Total Maximum Daily Load (TMDL) and EU Water Framework Directives (WFD), respectively. This dissertation aims to achieve water quality objectives using two reactionary approaches. First, I designed and assessed effectiveness of structural best management practices (BMPs), riparian buffers, to prevent agricultural nutrient and sediment loss from reaching streams. Second, I took a novel approach for the US and optimized agricultural system allocation across a watershed in order to achieve water quality improvements by maximizing the landscapes capability for the ecosystem services in water quality and food production. I hypothesized that adverse effects of agricultural activities on water quality can be minimized through comprehensive evaluations and flexible designs of riparian buffer BMPs and synthesized land use planning according to soil and landscape potentials or suitability. Chapter 2 presents results of the first approach. I examined the effectiveness of riparian buffers in reducing nutrient and sediment loads to streams as a function of buffer attributes and upland loads flowing into riparian zones in Spring Creek watershed in PA. Many agricultural watersheds in Pennsylvania rely heavily on the widespread adoption of buffers to meet water quality goals. However, landowners have identified design and management constraints in current policies as barriers to adoption. Therefore, this research explored water quality tradeoffs that may result from a more flexible buffer design paradigm. I hypothesized that a flexible buffer design paradigm, with variations in buffer vegetation and harvesting options, would not damage a buffer’s water quality benefits. I tested the effectiveness of four buffer designs (variations in vegetation and width), and two alternative buffer management scenarios, which involve harvesting of either grass or trees from the buffer. The buffer design and management scenarios were selected based on relevant riparian buffer policy and farmers’ buffer design preference. Three crop rotations were simulated in the Soil & Water Assessment Tool (SWAT) in terms of their nutrient and sediment loads, and which were further coupled to the Riparian Ecosystem Management Model (REMM), in order to better understand how input loads affect the effectiveness of a specific buffer design and how the effectiveness of a buffer design changes as a function of input load. Results revealed that grass, which is generally a farmer-preferred buffer vegetation, is equally or more effective in nutrient and sediment removal than trees. Results also suggest that harvesting part of the buffer, either trees or grasses (a farmer-preferred management practice too), had a marginal effect on a buffers’ water quality benefits. Results also show that farmer adoption recommendations should be made based on mass loads treated by the buffer instead of a buffer’s percentage reduction rates, as higher nutrient and sediment masses may be reduced by buffer with lower removal efficiencies when the buffer receives higher nutrient and sediment loads. Results from this study can help inform development of multi-function riparian buffer policies and help evaluate the impacts of such policy recommendations on stream water quality. Future research is suggested on comparing other ecosystem services that policy-recommended and farmer preferred buffers can provide. In Chapter 3, I explored the effect of riparian zone soil characteristics on riparian buffers’ effectiveness in the Upper Bann watershed in NI. NI experiences elevated phosphorus (P) concentrations in surface water bodies but has a very different style of agriculture from Spring Creek Watershed in PA and a very different soil and geomorphic model (glaciated wet landscape in NI versus the opposite in Spring Creek Watershed, PA). Riparian buffers in NI have been proposed as an environmental farming scheme to mitigate agricultural pollutants, such as phosphorus (P). The effectiveness of proposed riparian buffer designs in reducing phosphorus loss is uncertain, especially considering the heterogeneity of riparian zone physical characteristics. I used the SWAT model calibrated to P concentrations in streams, to simulate P losses from seven selected fields representative of the Upper Bann watershed. By incorporating SWAT outputs of nutrients and sediment loss into a riparian buffer model, REMM, the performance of five buffer designs (10 m grass, 10 m tree, 15 m grass, 15 m tree and 30 m grass and tree) were tested in 7 representative riparian buffer fields in the study watershed. Buffer performance was evaluated in terms of an annual reduction rate and an average reduction rate calculated for the top 25%, 10% and 5% of storm events. Results reveal that representative fields in the study watershed can be divided into four categories based on which buffer design works best for dissolved mineral P removal (at annual average scale) and what dissolved mineral P and total P (TP) reduction rates can be achieved. A soil’s hydraulic conductivity and water storage capacity are the key factors effecting buffer removal efficiency; well drained, deep soils are more suitable for buffer installation. In addition, the TP reduction rate, per a specific buffer, is generally lower than that of dissolved mineral P since non-active P (dissolved organic P) loads are increased by the buffer. Last, simulation results reveal that a buffer can export phosphorus to streams via surface and subsurface flows even when it does not receive phosphorus influx from the upland, and this especially occurs when soils of buffers are enriched with P before buffer installation. Results from this study demonstrate the importance of accounting for the environmental heterogeneity among fields in a watershed when evaluating buffer effectiveness in reducing nutrient export. This study also illustrates that a buffer can act as a phosphorus source even when it does not receive phosphorus influx from upland. Future research is suggested to verify the modeling results based on the field measurement and provide more accurate buffer installation recommendations based on riparian soil properties. In Chapter 4, I took a novel approach to nutrient management and optimized crop rotation patterns per landscape characteristics in order to minimize water quality deterioration in an agriculture watershed while maintaining crop productivity similarly to current baseline conditions. Common strategies to reduce agricultural pollution include a variety of structural best management practices (BMPs) and do not always lead to desired water quality improvement. I hypothesized that the spatial reallocation of agricultural commodities grown across a watershed may provide water quality benefits without adoption of new BMPs. I investigated how land use reallocation in the Conewago Creek Watershed in PA, without the introduction of additional BMPs, would change watershed water quality and crop yields. The Soil and Water Assessment Tool (SWAT), was utilized to model crop growth, hydrology, and loadings of nitrogen, phosphorus and sediment from 2010 through 2017 and to quantity the water quality regulation and food provision ecosystem service provided by different landscape characteristics. Optimization processes are conducted under two constrains: 1) the acreage of each crop rotation is remained the same as baseline to maintain crop yield; and 2) the reallocation only occurs within the current agricultural land. The optimization principle places crop rotations that generate the highest nutrients and sediment loss in areas that are least vulnerable to erosion and nutrient loss. SWAT simulation results reveal that crop reallocation alone was able to achieve a 15.0% total nitrogen (TN) reduction, a 14.4% total phosphorus (TP) reduction and a 39.0% sediment reduction at the annual average scale and maintain crop yields similar to current baseline conditions. However, further efforts are needed to understand how such watershed-level benefits may impact farm-level factors, as implementation would require farmers to change their crop land use type and re-distribute earned wealth across the watershed’s farms.