Linking Water Quality Trends with Land Treatment Trends

Linking Water Quality Trends with Land Treatment Trends
The Rural Clean Water Program Experience

Land use and land management affect the type and amount of nonpoint source (NPS) pollution entering water bodies. Improvements in land management (also referred to as land treatment) are necessary to reduce the delivery of pollutants to impaired or threatened water resources. Documentation of the magnitude of water quality improvements from changes in land management, for at least a few projects in each part of the country, is essential to provide feedback to project coordinators and state, regional, and national policy makers. Such feedback enhances the development and implementation of land treatment programs that effectively reduce delivery of pollutants causing water quality impairment. In addition, demonstration that land treatment is effective in reducing NPS pollution and improving water quality tends to increase political and economic support for NPS pollution control measures.

Historically, it has been difficult to demonstrate the relationship between land treatment and water quality changes, at least in part because of a lack of well-designed water quality and land treatment monitoring efforts. Two goals must guide the design of monitoring networks and data analysis in programs and projects designed to link water quality changes with implementation of best management practices (BMPs): 1) detection of significant (or real) trends in both water quality and land treatment and 2) linking or associating water quality trends with land treatment trends.

This fact sheet outlines the principles for development of effective monitoring designs, and describes the land treatment and water quality monitoring elements necessary for linking land treatment or land use modifications with water quality changes. These monitoring elements are essential for successful experimental watershed projects designed to document the relationship between land treatment and water quality changes.

Many of the recommendations for monitoring discussed in this fact sheet are based on the 15-year Rural Clean Water Program (RCWP), an experimental, agricultural watershed, NPS pollution control program that combined land treatment and water quality monitoring in a continuous feedback loop to document NPS control effectiveness (Gale et al., 1993; Spooner and Line, 1993).

Documenting a Cause-and-Effect Relationship

Documenting that water quality changes at a watershed scale were caused by implementation of BMPs is difficult. Not only must a strong correlation be established, but the observed changes must be repeatable over time and space in an experimental manner. The only major changes made in the watershed during the evaluation period should be changes in land treatment. Observed changes in water quality should match predicted pollutant reductions based on the estimated land treatment effectiveness. Some projects have been able to document a strong relationship, increasing confidence that appropriate land treatment can result in (cause) improved water quality. The stronger the relationship, the more likely it is that a cause-and-effect relationship exists and that water quality changes are caused by changes in land treatment rather than other factors.

An association (statistically significant correlation or relationship) between land treatment and water quality changes is required to demonstrate a cause-and-effect relationship. As the implementation of land treatment (specifically BMPs) occurs, improvements in water quality are observed. However, an association by itself is not sufficient to infer a cause-and-effect relationship. Other factors not related to BMP implementation may be causing the changes in water quality, such as changes in land use or rainfall. If, however, the association is consistent and responsive and has a mechanistic basis, causality may be supported (Mosteller and Tukey, 1977).

Consistency means that the relationship between the measured variables (such as total phosphorus and acres treated with the nutrient management BMP) holds in each data set in terms of direction and degree. A consistent, multi-year, improving trend in water quality after BMP implementation provides evidence needed to attribute water quality improvements to land treatment. Improvements in multiple watersheds treated with systems of BMPs provide strong evidence that water quality improvements resulted from land treatment.

Responsiveness signifies that as one variable changes in a known, experimental manner, the other variable changes similarly. For example, as the amount of land treatment increases, further reduction of pollutant delivery to the water resource is documented.

Mechanistic means that the observed water quality change is that which is expected based on the physical processes involved in the installed BMPs. For example, based on knowledge of absorption and solubility of nutrients, greater reduction of nutrient delivery to the water resource might be predicted as the result of implementation of the manure management BMP than a soil erosion control practice alone.

Elements of Monitoring Needed to Link Land Management Modifications with Water Quality Changes

Experimental Designs for Water Quality and Land Treatment

An appropriate experimental design for water quality and land treatment monitoring is essential to document a clear relationship between land treatment and water quality changes. The best designs to demonstrate linkage are those that can isolate the effects of the land treatment from other land use and climatic changes. Such designs include: 1) paired watershed (Clausen and Spooner, 1993); 2) upstream-downstream sites monitored before, during, and after land treatment; and 3) multiple watershed monitoring.

The paired watershed design is the best method for documenting BMP effectiveness in a limited number of years (three to five). Two or more similar subwatersheds (drainage areas) are monitored before and after implementation of BMPs in one of the subwatersheds (the treatment subwatershed). Paired drainage areas should have similar precipitation and runoff patterns and should exhibit a consistent relationship in terms of the magnitude of pollutant losses with changes in hydrology and climate. Analysis of paired pollutant data from treatment vs. control areas should show a statistically significant correlation. Ideally, a paired watershed monitoring program is characterized by:

Simultaneous monitoring at the outlet of each drainage area;
Monitoring prior to land treatment to record the relative hydrologic response of each drainage area (calibration period);
Calibration period of one to three years, depending on the consistency of the data relationships between drainage areas;
Subsequent monitoring where at least one drainage area continues to serve as a control (that is, receives significantly less land treatment than the other drainage area); and
Similar land management in both drainage areas both before and after BMP implementation (for example, similar crops), except for BMPs implemented in the treatment drainage area.

Land Management and Water Quality Monitoring Before and After BMP Implementation

Monitoring for several years both before and after BMP implementation is essential for documentation of water quality changes. The pre-BMP period is the time prior to installation of new land treatment practices. Monitoring of water quality and land use prior to BMP implementation is required to establish baseline data for statistical comparison with post-implementation data. The post-BMP period starts once BMPs have been implemented on critical areas and are reducing pollutant delivery to the water resource.

Sampling frequency and collection must be consistent across seasons and years. Year-to-year variability is often so large that at least two to three years each of pre- and post-implementation monitoring is required to indicate a consistent water quality change following implementation and maintenance of BMPs. Documentation of changes over multiple years increases confidence that observed water quality improvements are due to land treatment.

Short-term monitoring is seldom effective because climatic and hydrologic variability can mask water quality changes. However, in small watersheds affected by relatively few large pollutant sources, the monitoring period may be shorter. Longer duration monitoring is necessary where water quality changes are likely to occur gradually, such as large watersheds with lakes in which lag times may occur due to buffering effects of long hydraulic residence times and pollutant recycling.

Quantitative Monitoring of Land Management

The importance of recording the amount and type of land treatment cannot be overlooked when trying to establish documented water quality improvements. Best management practices must be targeted to treat specific sources of pollutants causing the water quality impairment; these pollutants, in turn, must be monitored in the water resource. A high level of appropriate NPS pollution control implementation in critical areas is usually required to achieve substantial water quality improvements.

Monitoring of land treatment and land use is needed to quantify the pollutant reduction impacts of BMPs. Quantitative monitoring of BMP implementation facilitates documentation of land treatment trends and is a necessary step in linking water quality to land treatment. Methods of reporting and quantifying land treatment and land use should be consistent throughout a project.

Careful planning is required to determine which land treatment variables should be monitored to best reflect the extent of actual changes in agricultural practices. Land treatment data must be reported in quantitative units that reflect BMP effectiveness and changes from previous practices. Examples of quantitative units include: application method, tons of manure spread per acre, pounds of fertilizer applied per acre, acres served by each BMP, and acres served by each BMP system. The acres served unit includes all treated acres (those acres with actual implementation) plus all acres whose pollutant delivery is being reduced by the BMP. Documenting the assumptions used in calculating the acres served is important so that these units can be calculated consistently from year to year, thus ensuring valid year-to-year comparisons.

When reporting acres served, care should be taken to avoid double counting acres when multiple BMPs are serving the same acres, as this could artificially inflate the reported number of acres served. In addition, correction should be given for differences in the effectiveness of the BMPs in controlling pollutant delivery.

Operation, management, and maintenance of BMPs should be tracked because these factors affect BMP effectiveness and, therefore, the water quality impacts of the land treatment.

Changes in land use should be recorded to help isolate the water quality changes associated with the NPS controls from water quality changes due to other land use factors. Land use modifications that affect water quality include acres converted from row crops to pasture (permanently or based on rotation), set-aside acres, changes in the number of animals or animal units per acre, closure of animal operations, changes in impervious land areas, implementation of soil and water conservation practices not being recorded as part of the project, and changes in non-agricultural land uses.

Matching of Land Treatment and Water Quality Data on a Spatial (Drainage) Scale

Land treatment data must be collected on a hydrologic or drainage basis so that the land area being tracked corresponds to the drainage area served by each water quality monitoring station. Water quality and land treatment data must be matched if water quality changes are to be attributed to BMP implementation.

Linkage of land treatment and water quality impacts can be made at different spatial scales (such as farm field, subwatershed, or watershed). Spatial scale should be determined based on project goals and desired interpretations. In general, the larger the drainage area, the harder it is to identify and quantify a water quality - land treatment linkage. Water quality changes are more likely to be observed at the subwatershed than watershed level. Confounding effects of external factors, other pollutant sources, and scattered BMP implementation are minimized at the subwatershed level. If the goal is to document changes at the entire watershed level, a monitoring station must be located at the watershed outlet.

Matching of Land Treatment and Water Quality Data on a Temporal Scale

Water quality and land treatment data should be collected during the same time periods so both data sets are temporally related. Actual implementation of land treatment needs to be recorded at least seasonally or annually. Land treatment data (such as timing of manure or commercial fertilizer applications, construction of a lagoon storage structure, or a dairy closure) should be collected more frequently than annually or seasonally if the effect on water quality is more short-term or has a large, immediate impact.

Water quality samples are usually collected weekly or biweekly. These data do not have to be summarized on the same time scale as the land treatment data; land treatment data can be added to the trend analysis as repeating explanatory variables. Alternatively, water quality data can be aggregated to the same time scale as the land treatment data for analysis. Data aggregation is particularly useful for plotting and explanatory data analysis.

Matching Monitored Pollutants with Pollutants Addressed by Land Treatment

Pollutants monitored at water quality stations must correspond to pollutant(s) being treated by the BMP systems implemented.

Monitoring Explanatory Variables

Accounting for all major sources of variability in water quality and land treatment data increases the likelihood of isolating water quality trends resulting from BMPs. Correlation of water quality and land treatment changes by itself is not sufficient to infer causal relationships. Other factors not related to BMPs may be causing water quality changes, such as changes in animal numbers, cropping patterns, land uses, known pollutant sources, or amount of impervious land surface; season; stream discharge; precipitation; ground water table depth; salinity; or other climatic or hydrologic variables. Factoring explanatory variables into trend analyses yields water quality trends closer to those that would have been measured had no changes in climatic or other explanatory variables occurred over time. Accounting for variability in water quality due to known causes also decreases variation in adjusted water quality data, facilitating documentation of statistically significant trends. Explanatory variables should be monitored at the same frequency as the principle water quality variables.

Summary

A good experimental design for water quality and land treatment monitoring is essential in order to provide clear documentation of the relationship between land treatment and water quality changes. The paired watershed monitoring design can best demonstrate the relationship between land treatment and water quality in the shortest period of time.

To determine if the trends in water quality match the mechanistic prediction of trends, pre- and post-BMP implementation monitoring and data analysis must combine water quality, land treatment, and land use data on suitable spatial and temporal scales. Incorporation of explanatory variables facilitates isolation of water quality changes that result from land treatment.

References

Clausen, J.C. and J. Spooner. 1993. Paired Watershed Study Design. Office of Water, U.S. Environmental Protection Agency, Washington, DC. EPA 841-F-93-009. 8 p.

Gale, J.A., D.E. Line, D.L. Osmond, S.W. Coffey, J. Spooner, J.A. Arnold, T.J. Hoban, and R.C. Wimberley. 1993. Evaluation of the Experimental Rural Clean Water Program. National Water Quality Evaluation Project, NCSU Water Quality Group, Biological and Agricultural Engineering Department, North Carolina State University, Raleigh, NC, EPA-841-R-93-005. 559 p.

Mosteller, F. and J.W. Tukey. 1977. Data Analysis and Regression: Second Course in Statistics. Addison-Wesley Pub. Co., Reading, MA. 588 p.

Spooner, J. and D.E. Line. 1993. Effective Monitoring Strategies for Demonstrating Water Quality Changes from Nonpoint Source Controls on a Watershed Scale. Water Science Technology, 28(3-5):143-148.

Written by

Jean Spooner, Daniel E. Line, Steven W. Coffey, Deanna L. Osmond, and Judith A. Gale

Water Quality Extension Specialists

NCSU Water Quality Group

March 1995

North Carolina
Cooperative Extension Service

NORTH CAROLINA STATE UNIVERSITY
COLLEGE OF AGRICULTURAL & LIFE SCIENCES

Distributed in furtherance of the Acts of Congress of May 8 and June 30, 1914. Employment and program opportunities are offered to all people regardless of race, color, national origin, sex, age, or disability. North Carolina State University, North Carolina A&T State University, U.S. Department of Agriculture, and local governments cooperating.

This fact sheet is one of a series of Rural Clean Water Program Technology Transfer fact sheets prepared by the NCSU Water Quality Group with support from the Extension Service, U.S. Department of Agriculture (Cooperative Agreement No. 93-EXCA-3-0241).

Copies of the fact sheet series may be requested from: Publications, NCSU Water Quality Group, Department of Biological and Agricultural Engineering, Box 7637, North Carolina State University, Raleigh, NC 27695-7637, Email: wq_puborder@ncsu.edu, Fax: 919-515-7448.