This is the first of seven chapters that review the individual programs of the Watershed Protection Program, including the Watershed Agricultural Program (WAP), the Stream Management Program, the Land Management Program, wastewater programs, the Stormwater Program, ecosystem protection programs, and public health programs. Each program is first described, including its goals, funding, and scope or geographic extent. Then, the effectiveness of the program is analyzed using a combination of data provided by the New York City Department of Environmental Protection (NYC DEP) or its program partners and analysis performed by the Committee. Finally, the program is broadly critiqued, ending with conclusions and recommendations for improving the program.
The WAP is a City-funded program that works with farm and forest landowners located in the Croton and Catskill/Delaware watersheds to protect water quality. The Watershed Agricultural Council (WAC) is a council of relevant watershed stakeholders, funded under a contract with NYC, that administers the WAP. WAC members include Delaware County Soil and Water Conservation District, Cornell Cooperative Extension, the U.S. Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS) and Farm Service Agency, NYC DEP, and watershed producers. WAC has two primary missions: to protect water quality (with a focus on nutrients, microbial pathogens, sediment, and pesticides) and to ensure the economic vitality of watershed agriculture. The mechanisms WAC uses to meet water quality goals include Whole Farms Plans, forest management plans, and conservation easements, of which forest management plans and conservation easements are covered in Chapters 10 and 7, respectively. WAC provides technical assistance to producers and educational materials for stakeholders. The entire WAP is voluntary.
In 1990, draft Watershed Rules and Regulations were presented to upstate watershed communities including language that would potentially negatively affect agriculture. Following this draft regulation, in 1991 an ad hoc task force on agriculture and the NYC Watershed Rules and Regulations convened, representing upstate and downstate interests. The so-called Town Brook Plan developed a set of policy recommendations to both protect water and allow continued agricultural operations in the watershed. Important recommendations from the Town Brook Plan included the withdrawal of the City’s proposed agricultural regulations and the development of the voluntary WAP funded by the City. The WAP, described below according to Graff (2018), formed the basis for the 1997 Memorandum of Agreement (MOA) between the NYC DEP and upstate watershed communities.
Phase I of the WAP began in 1992, with the creation of Whole Farms Plans on ten pilot farms in the Catskill/Delaware watersheds. In 1993, the partners decide to create the non-profit WAC to administer the
program, fully funded by NYC. In 1994, Phase II of the WAP began with the development of Whole Farms Plans for watershed farms based on USDA standards. A goal of 85 percent participation of all large farms (>$10,000 revenue) became a milestone in the U. S. Environmental Protection Agency’s (EPA’s) filtration avoidance determination (FAD) waiver, a goal that was met and exceeded as the WAP matured and expanded over the course of two decades. Today WAP no longer differentiates between large and small farms, and currently has Whole Farm Plans in place for 331 eligible farms, or 74 percent of all farms, and 70.7 percent of all agricultural acreage. In 1996, the Watershed Forestry Program was created, followed in 1998 by the Conservation Easement Program, implemented to purchase the development rights on watershed farms.
Funding and Geographic Extent
Since the WAP’s creation in 1993, the City has invested $212,876,000 to support the program. Annual best management practice (BMP) implementation budgets typically range from $2 million to $3 million and have been as high as $5 million (Figure 5-1). This total of about $212 million includes over $61,000,000 spent on BMPs, including 6,829 BMPs on west-of-Hudson (WOH) farms ($54.7 million) and 734 BMPs on east-of-Hudson (EOH) farms ($6.5 million). The current BMP backlog is about $41 million, which includes 1,410 new BMP installations (estimated cost of $28.1 million) and 344 BMP repairs or replacements (estimated cost of $7.7 million). During 2017, WAC reduced the backlog by 75 new BMPs (5 percent) and 94 repair/replacement BMPs or 27 percent (NYC DEP, 2019).
NYC DEP is the primary source of BMP funding, with the only other source being the Conservation Reserve Enhancement Program (CREP). CREP is a federal program, administered by the USDA Farm Service Agency, to exclude livestock from streams and establish or restore riparian forest buffers, grass and shrub buffers, and wetlands via land rental and cost-share payments to farmers who retire near-stream areas from agriculture. All CREP-enrolled pasture or cropland is planted with hardwood trees, native warm season grasses, or approved shrubs and grasses. The program funds restoration of wetlands, forested riparian buffers up to 300 feet wide, and cropland buffers up to 100 feet wide. WAC does not leverage this USDA cost share for any other BMPs.
WAC works with farm and forest landowners in the Croton and Catskill/Delaware watersheds. However, most participants in the WAP are in the WOH basins, where 264 active farms are supported (a participation rate of 71 percent). Since the activation of the water filtration plant for the Croton water supply system, NYC
DEP has deemed the watersheds that feed directly into the Croton system as low priority for BMP funding and admittance into the WAP.
Whole Farm Plans
WAP addresses water quality through implementation of BMPs and land conservation as captured in a Whole Farm Plan. These BMPs, along with conservation easements, are developed for the individual farm property. As is evident from the prioritization list below, WAP places the highest priority on storage of animal waste, which is a source of pathogens and nutrients, including nitrogen, phosphorus, and organic carbon. WAP also considers numerous other risks to the water supply including pesticides and sediment.
Whole Farm Plans are contracts administered between WAP and participating producers that identify and prioritize environmental issues on farms. Potential risks to the water supply are identified and addressed through structural planning to reduce or prevent the transport of agricultural runoff and contaminants to waterbodies. The process begins when a farmer signs a voluntary participation agreement with WAP and agrees to develop a Whole Farm Plan in conjunction with a WAC Planning and Implementation Team.
To select BMPs, the Planning and Implementation Team visits the farm to identify and assess potential sources of pollutants, using an Environmental Review/Problem Diagnostics tool. The team reviews technical and financial options with the landowner/farmer and drafts a Whole Farm Plan, which is presented to WAP peers and managers for review. The landowner/farmer signs the Participation Funding Agreement and Whole Farm Plan agreement to implement their plan, following the BMP prioritization process and subject to WAP funding availability. The final Whole Farm Plan is presented to the WAC Agricultural Program Committee for approval.
Given the popularity of the program among watershed farmers and the substantial needs on many individual farms, it has been a challenge for WAC to prioritize BMP installations within and among farms. Therefore, in 2011 WAC adopted a BMP Prioritization Methodology to establish implementation priorities among individual farms (all farms are ranked based on risk of nutrients and pathogens moving off the farm) and within each Whole Farm Plan (all BMPs are ranked based on priority pollutant categories within each farm). To rank farms based on risk, the factors listed in Table 5-1 are considered. Each factor has a zero score level and a max score level, and factors are weighted (between the zero and the max level) by linear interpolation.
Second, at individual farms BMPs are selected and implemented by using a source categorization process. Potential pollutants are categorized and prioritized as follows:
- Parasites & Phosphorus: Animal Waste Storage
- Pesticides: Storage Facilities, Mixing/Loading Areas
- Phosphorus: Fertilizer Storage
- Parasites: Animal & Manure Management
- Nutrient Management/CREP Buffers
- Nutrients: Concentrated Sources
- Sediment: Diffuse
- Sediment: Concentrated
- Pesticides: Field & Animal Application
- Fuel Storage
- Other Materials
Despite this list, in reality, after the farm rankings have been determined, CREP riparian buffers and nutrient management plan BMPs (Category V, above) are implemented first, starting with the first-ranked farm and repeating for all farms in ranked order. The reason that CREP buffers are implemented first is because most of the cost share in the CREP program is from federal sources, making these low-cost BMPs for WAP. A total of 1,843 acres of CREP riparian forest buffers are currently enrolled in 184 active contracts representing 141 different landowners.
TABLE 5-1 Farm Ranking Score Factors, Data Requirements, and Weighting Levels
|Factor||Description||Max Level||Zero Level||Max Points|
|Animal density (AU/acre)||Animal unit (AU): number of 1,000-lb animal equivalents on the farm, from annual status report (ASR)||1 AU/ac||0 AU/ac||20|
|Acres: total of hayland, cropland, and pasture acres on farm, from Nutrient Management Plan (NMP)a|
|Farm size (total AU)||AU: number of 1,000-lb animal equivalents on the farm, from ASR||100 AU||5 AU||20|
|Young stock||Number of potentially Cryptosporidium-shedding animals < 12 months old, from ASR||20 animals||0 animals||20|
|Soil test phosphorus (% acres with very high P)||Percentage of acres on the farm with a soil test P of greater than 39 lb/ac, from NMP||35%||0%||10|
|Field distance to watercourse||Average of the distance to a watercourse for all fields on farm, from NMP||50 ft||500 ft||10|
|Farmstead proximity to watercourse||Distance of primary livestock housing location to a watercourse,b from planner||300 ft||1,000 ft||20|
a Nutrient management plans document all crop nutrient needs, soil test results, and application of all nutrients (including manure) to the fields in an attempt to balance soil nutrient inputs with crop requirements.
b The distance (in feet) of the primary livestock housing location (barn) to a watercourse visible on GIS aerial photography, rounded to the nearest 25 feet. A watercourse is a visible path through which surface water travels on a regular basis, including an intermittent stream. A drainage ditch, swale, or surface feature that contains water only during and immediately after a rainstorm or a snowmelt is not considered a watercourse.
SOURCE: Adapted from Graff (2018).
Next, BMPs intended to deal with pollutant categories I-III are implemented, again following the farm rank order. After that, the BMP list is sorted by farm to develop a prioritized list for each farm area, and BMPs are selected annually from each farm list for installation based on the budget allocation.
Once a Whole Farm Plan is completed and BMPs are operational, participants receive and agree to a BMP Operations and Maintenance (O&M) Agreement. This O&M agreement is a long-term contract that outlines the steps a landowner takes to maintain the Whole Farm Plan infrastructure. The O&M agreement lasts the lifespan of a BMP, usually 10 to 15 years.
A WAC staff planner conducts an Annual Status Review of the BMPs, their functionality, and effectiveness, which includes:
- A WAC planner/landowner discussion about how the Whole Farm Plan worked over the past year and identification of changes made on the farm in the last year that may impact the Whole Farm Plan in the coming year;
- BMP inspections to verify they are in place and maintained;
- Certification that the landowner is following the O&M agreement;
- Scheduling pending BMP installations; and
- Measuring landowner satisfaction with the Whole Farm Plan and its overall impact on farm operations.
Whole Farm Plan revisions occur when the farm’s water quality issues are not sufficiently addressed by the existing plan and may be due to changes in the farm business, the physical nature of the farm (such as erosion caused by a flood event), or a failure of an existing BMP to adequately address an original issue. The revision process is similar to the original Whole Farm Plan process.
Nutrient Management Plans
Nutrient Management Plans (Category V above) are a central component of the Whole Farm Plan process, intended to assist producers in managing agricultural nutrients. Specifically, a Nutrient Management Plan details the amount, source, placement, form, and timing of the application of plant nutrients (manures and fertilizers) and soil amendments. These plans are rewritten every three years by WAC and include field-level soil phosphorus and manure phosphorus tests, an evaluation of field connectedness to surface waters using a flow-path analysis, and annual soil loss estimates using the USDA-NRCS Revised Universal Soil Loss Equation 2. A WAC planner works with a producer to develop a nutrient budget for a given field. Data used to develop the nutrient budget include information about crop rotation, soil and tissue nutrient content, and realistic crop yields. The prescribed rates of phosphorus application, as documented in the nutrient budget, are based on guidance from the NYS NRCS 590 standard, which is summarized below.
Phosphorus application rates to a field should match rates recommended for a given crop provided by the Cornell Cooperative Extension as closely as possible. Phosphorus-based manure application rates are determined as a function of soil test recommendations or estimated phosphorus removal in harvested crop biomass based on guidance from the NRCS General Manual (NRCS, 2012, Title 190, Ecological Sciences, Part 402, Nutrient Management), and the National Agronomy Manual (NRCS, 2011, Part 503). The nutrient content of manure should be determined prior to land application, ideally annually. For manure storage, the nutrient content should be analyzed with each emptying cycle. Phosphorus application should also be consistent with the New York Phosphorus Index rating. For sites at low or medium risk for phosphorus loss (as determined by the Phosphorus Index score), the manure application rate is based on nitrogen requirements of the crop. For high-risk sites, the manure application rate is based on phosphorus removal of the crop. On very high-risk sites, no manure application is allowed.
Farmers can receive a Nutrient Management Credit as an incentive for following the Nutrient Management Plan. A peer-review committee reviews producer compliance with the plan and approves the credit. A credit of $10 per acre and $11 per animal can be used for expenses related to nutrient management (e.g., manure storage, manure spreaders, expenses related to trucking manure).
Precision Feed Management Program
The Precision Feed Management Program, initiated in 2016, attempts to reduce the nitrogen and phosphorus content in livestock feed, ultimately resulting in less nitrogen and phosphorus in manure excretions. There are currently 54,985 kg of phosphorus and 360,386 kg of nitrogen under program management in the WOH system (primarily in the Cannonsville), responsible for 37,761 kg of phosphorus and 263,003 kg of mitrogen in manure excretions annually (with the difference being incorporated into animal biomass, milk, meat, etc.). Currently, 46 farms are participating in the Precision Feed Management Program. Using a pseudo-randomized design (various levels of feed nitrogen and phosphorus adjustments, and a control), the program has been able to quantify significant reductions in phosphorus excretions in 2016, 2017, and 2018. Phosphorus excretions in manure declined by 4.7 kg/cow/yr (2,795 kg/yr, a 26 percent reduction) in 2016, by 5.5 kg/cow/yr (2,470 kg/yr, a 23 percent reduction) in 2017, and by 8.4 kg/cow/yr (2,905 kg/yr, a 31.5 percent reduction) in 2018. These are substantial per cow reductions; however given the number of farms participating, they are a small portion of the overall total phosphorus flux to the Cannonsville Reservoir of approximately 78,000 kg/yr (see Chapter 4). If participation in the Precision Feed Management Program were to expand to include a greater proportion of WOH farms, this program could result in significant nutrient reductions and begin to bring the nutrient mass balance back into equilibrium.
In 2018, WAP instituted two new initiatives—aerial cover crop seeding and the Precision Feed Management pilot lime project. Cover crops are typically grass or leguminous crops planted after primary crop harvest (usually in the fall) and are intended to provide soil cover, to prevent erosion, and to perform some nutrient uptake or fixation (for the legumes). Cover crops are not typically harvested but are killed down prior to the next season planting. The aerial cover crop seeding initiative led to a substantial increase in acres of cover crops seeded, from 50 to 875 acres. This resulted in an estimated 10,700 kg of N temporally prevented from being exported.
The second new initiative, the pilot lime project, was instituted to address low pH farm soils, which reduce crop growth and crop uptake of nutrients, and can increase nutrient loss. Using lime as a soil amendment, this project was implemented on 15 Precision Feed Management-participating farms with the goal of raising soil pH into the range of 6.2-7.0, which is more amenable to growth.
Of critical importance to the Watershed Protection Program is the extent to which its individual programs address the overarching goals of the MOA to “maintain and enhance the quality of the New York City drinking water supply system and the economic vitality and social character of the Watershed communities.” The WAP is typical of most of the Watershed Protection Programs in using a variety of metrics and methods to understand the effectiveness of the program, and yet it employs no direct measurements of water quality or community vitality. Thus, WAC does not directly conduct water quality monitoring at farms, although they have funded special studies that do. Rather, the program relies on measures, such as the number of participants, plans, and BMPs, along with annual status reviews of Whole Farm Plans that track changes in farm operation (i.e., animal units, acreage, newly identified pollutant issues, BMP issues, and other landowner concerns).
For farms with a Nutrient Management Plan soil phosphorus is measured to assess the need for commercial fertilizers and to manage manure spreading. WAC has not done an in-depth study of field-specific nutrient changes over time, although a field-by-field soil sample database does exist going back to 2010. The Committee’s analysis of these soil phosphorus data is discussed below.
Over the last 20 years, special studies have been done to assess the effectiveness of BMPs in reducing pollutant loading to waterbodies. In 1993 WAC funded a comprehensive, long-term paired watershed basin study (from 1993 to 2005; Bishop et al., 2005), which assessed phosphorus reduction from aggressive BMP implementation. Data from this study, which evaluated water quality on one dairy farm pre- and post-BMP implementation, have been used in subsequent modeling efforts to estimate phosphorus reduction from BMPs using the Soil and Water Assessment Tool (SWAT, Cerucci and Pacenka, 2003), the Generalized Watershed Loading Function (GWLF) model (NYC DEP, 2006), and the Variable Source Loading Function (VSLF) model (Easton et al., 2008a). The relevant study results are discussed in the section on Paired Watershed and Other Case Studies. Although the WAP does not support collection of BMP performance data related to pathogens, two studies looked at BMP effectiveness in reducing pathogen risk on farms: Use of a Sentinel System for Field Measurements of Cryptosporidium parvum Oocyst Inactivation in Soil and Animal Waste (Jenkins et al., 1999) and Evaluation of Solar Calf Housing for the Presence and Survival of Cryptosporidium parvum in the NYC Watershed (Collick, 2003).
Current Program Metrics
WAC uses several metrics to evaluate and judge the success of the overall program. Most metrics are based on compliance with the FAD and MOA, and generally include evidence of the number of people participating in the programs or the number of BMPs installed. However, many additional relevant metrics could be employed
to quantify performance of the program in improving water quality. Both types of metrics are discussed in the following section. Data and metrics reported in this section can be found in NYC DEP (2019), WAC (2016, 2017), and WAP (2018).
Participation Rate in WFP Process
The FAD mandates a 90 percent participation rate of large farms (defined as gross revenues >$10,000/yr) as of 2016, (which the WAP has met in each evaluation year). WAC has developed 449 Whole Farms Plans on 371 WOH farms and 78 EOH farms. 71 percent of all WOH farms (264, includes large and small operations) and 86 percent of all EOH farms (67, includes large and small operations) remain active. Total acreage for all active program participants is estimated at 151,000 acres (70.7 percent of the total agricultural land), which includes nearly 44,000 acres of cropland or pasture (61.2 percent of all crop or pasture land) actively managed through Whole Farms Plans.
Of the 264 active WOH farms, 176 are large and 88 are small. 62 eligible farm operations, representing 1,006 animal units (AU), are not participating in the WAP. Participation in the program is voluntary, yet the Surface Water Treatment Rule (SWTR) criteria for watershed control favors complete watershed land ownership or agreements with all landowners—focused on activities or areas with potential to affect water quality. A very high degree of control/mitigation in some areas may balance out risks from other less-controlled areas (assuming that high-control areas and low-control areas produce a similar water quality response). If areas under control present a greater or lower risk to water quality, then determining the effectiveness of such a program becomes more difficult and points toward the need for a strategic evaluation and prioritization, which can be difficult, expensive, and time-consuming with a voluntary program.
Number of Annual Status Reviews Conducted
Another metric is the completion rate of annual status reviews on at least 90 percent of all active Whole Farm Plans. Although the number of status reviews required each year varies based on the number of active Whole Farm Plans, the WAC completed an average of 318 status reviews each year on large and small WOH and EOH farms; this equates to an average annual completion rate of 94 percent. Annual status reviews track changes in animal units and acreage, newly identified pollutant issues, and BMP O&M issues, as well as any concerns the landowner may have with their Whole Farm Plan, but does not directly track water quality benefits.
Number of Nutrient Management Plans Enrolled
Another program metric is to maintain Nutrient Management Plans on 90 percent of all participating large farms. During the last five years, WAC maintained current Nutrient Management Plans on 94 to 100 percent of all participating large farms, with the actual number of farms needing these plans fluctuating between 174 and 180 farms annually. WAC will continue to make the Nutrient Management Credit Program (discussed earlier) available to at least 100 watershed farms. During the current FAD assessment period, the number of farms receiving nutrient management credits increased every year, from 91 farms receiving credits in 2011 to nearly 120 farms receiving credits in 2015. More than 100 farms per year have received nutrient management credits since 2012.
Number of BMPs Installed and Maintained
WAC has installed 7,565 BMPs (6,829 in the WOH watershed, totaling $55,000,000, and 734 in the EOH watershed, totaling $7,000,000), which equates to $182,034 spent on BMPs per farm enrolled (approximately $700,000 per large farm). As mentioned earlier, WAC adopted a new framework in 2011 for scheduling and
implementing BMPs across all participating farms in a manner that they suggest provides the greatest protection to water quality. During 2011-2015, WAC implemented 992 new BMPs totaling $4 million, in addition to repairing or replacing 246 failing or outdated BMPs totaling $2.4 million.
Number of CREP Participants
Developing new and reenrolling expiring CREP contracts is a metric used by WAC. During the current FAD assessment period, 26 new contracts consisting of 164.7 acres of riparian buffers were enrolled in CREP, while 29 expiring contracts consisting of 301.9 acres were reenrolled. Twenty-one (21) expiring contracts consisting of 136 acres were not reenrolled by choice of the landowners. As of December 2018, there were 179 active CREP contracts with 136 different landowners representing 1,809.9 acres of riparian buffers.
Number of Farmers Educated
The numbers of Farmer Education and Farm-to-Market Programs is used as a metric of program success. During 2011-2015, WAC conducted more than 140 farmer education programs that were attended by over 3,400 participants, of which 44 percent were watershed farmers, with the remainder being watershed residents or interested parties.
Other Metrics: Committee’s Analysis of Soil Phosphorus Data and Watershed Study Results
To determine if Nutrient Management Plans are resulting in soil phosphorus reductions, the Committee conducted a statistical analysis of the trends in soil phosphorus1 levels over the period 2010 through 2018. The data set analyzed consisted of 10,041 soil phosphorus values, collected from a total of 3,529 fields in the Cannonsville Reservoir watershed (data supplied by Dale Dewing, WAC, personal communication, October 2019). The number of years in which phosphorus measurements were made on any given field ranged from as few as one to as many as five years with a median of three years of sampling per field. These data were collected for operational purposes by WAC, to provide guidance to farmers on the need for applying phosphorus to their fields. Although the data set is not ideal for research, it can provide a general impression of trends in soil phosphorus across the watershed. By themselves, data from an individual field almost never give statistically significant trends (primarily due to too few measurements from individual fields), but when aggregated across the whole watershed or by various categories, it can provide strong evidence of trends. This analysis is accomplished by using the Regional Kendall test (Helsel and Frans, 2006), which helps aggregate trend results across multiple sites. This nonparametric method is used because the data set shows substantial skew and high outliers. The full data set is shown in the boxplot in Figure 5-2.
Soil Phosphorus Levels by Crop Type
Before describing the trend analysis, it is informative to graphically characterize the soil phosphorus values throughout the watershed. First, the Committee explored average soil phosphorus values as a function of crop type. The results appear in Figure 5-3. Of the three major crop types (hay, pasture, and crop rotation), the crop
1 Soil phosphorus concentration classifications for the Morgan Soil Test phosphorus level used by WAC are divided into four classes: low (<5 lbs/ac), medium (5-9 lbs/ac), high (9-39 lbs/ac), and very high (>39 lbs/ac) phosphorus levels. Fields with soil-test phosphorus levels in the low and medium classes can be managed using nitrogen-based fertilizer/manure application rates (e.g., nutrient applications need not consider the phosphorus content of the fertilizer or manure). Fields with high soil-test phosphorus levels are required to utilize phosphorus-based fertilizer/manure application rates (e.g., phosphorus applications cannot exceed plant uptake). Fields with very high soil-test phosphorus levels should receive no phosphorus application.
rotation category had the highest soil phosphorus values, with more than 50 percent of the measured values being above the high soil phosphorus threshold (>9 lbs/ac). The other two major crop types (hay and pasture) had more than 25 percent of their values above this threshold, but their median values were about a third as high as the crop rotation category. The vegetable category, although a small category, showed the highest soil phosphorus values, with more than 75 percent of the fields in the high category and more than 25 percent of the fields in the very high category (>39 lbs/ac). These results suggest that the greatest focus of efforts at drawing down legacy soil phosphorus should focus on crop rotation and vegetable fields.
Soil Phosphorus Levels by Soil Wetness Index Class
For this analysis, each field was divided into pixels of 10 m by 10 m. Each pixel was characterized as belonging to one of ten soil wetness index classes derived from a topographic index (TI) described later in this chapter and shown in Figure 5-4, with class 1 being the least likely to be saturated and class 10 being most likely to be saturated and generate surface runoff. Higher soil wetness index classes tend to be characterized by several properties that generate greater runoff losses, including large upslope drainage area, shallow slopes, and/or low soil transmissivity. For each field the median TI value was calculated from all the pixels in that field (note that a given field might contain pixels that include the entire range of soil wetness index classes or only a small subset of soil wetness index classes). The mean soil phosphorus value was computed for each field (regardless of what year it was sampled) and boxplots of these mean values are shown as a function of the median soil wetness index class for that field (shown in Figure 5-5). These results show that soil phosphorus values tend to rise with increasing soil wetness index class and are particularly high in the two wettest classes. Soil wetness index class 9 has 48 percent of its soil phosphorus values in the high range and 11 percent in the very high range. Soil wetness index class 10 has 60 percent in the high range and 12 percent in the very high range. These results suggest the need for focusing the efforts to draw down legacy phosphorus in areas of these wetter, more runoff-prone soils.
Trends in Soil Phosphorus Over Time
The Committee also explored changes in soil phosphorus values over the nine-year period. Figure 5-6 shows changes over time but is not a formal trend test. It suggests that there has been some decline in phosphorus values, but with considerable interannual variability. The characteristics of the fields sampled may have varied by year and this confounds the ability to see a temporal trend with these data (but see analysis in
the following paragraphs, which resolves this issue). Considering the first three years together, the percentage of the soil phosphorus values in the high range is approximately 40 percent and in the last three years of the record it is about 35 percent.
Trends in soil test phosphorus values can be effectively evaluated, even with variable datasets, when appropriate statistical tests are used. In this case, for all fields where at least two years of soil test phosphorus data are available, the Regional Kendall test statistic is computed, aggregating trend information from each of the 3,529 fields (see Appendix A for a description of the calculation method). The test is nonparametric (and thus resistant to the influence of extreme outliers) and it uses “blocking” which means that values are only compared with other values collected from the same field. The blocking removes the confounding effects of differences among fields. The result for this test is a statistically significant downward trend in soil test phosphorus (two-sided p-value < 0.001). The slope of that downward trend is -2.6 percent per year. This result, along with results for various subsets of the full data set (based on crop type, soil wetness index class, and field size), is shown in Table 5-2.
The results were further categorized by the crop type: hay, pasture, crop rotation, vegetables, and other crops. All five of these crop types showed tendencies toward decreasing soil phosphorus values over time and these decreases were significant in the case of the three major crop types (hay, pasture, and crop rotation). The other two crop types had far fewer fields sampled and their trends were not statistically significant (probably a result of the small sample size). The reductions in soil phosphorus values for all crop types tended to be small in percentage terms. If the trends were expressed as a half-life, a trend of -2.0 percent/yr would translate to about 34 years to reach a phosphorus value of half of the value observed at the first year of sampling. The point here is that the declines appear to be real for the major crop types, but the decline is relatively slow, as noted in other studies of soil phosphorus drawdown (Fiorellino et al., 2017).
The fields were also categorized by the average phosphorus levels observed over the nine years sampled. Trends in soil phosphorus levels did not quantitatively differ among categories of different initial soil phosphorus levels. That is, percentage change in phosphorus levels were similar across the range from high-phosphorus fields to low-phosphorus fields. We also categorized the results by field size because larger fields with high soil phosphorus and or high soil wetness are a large potential source of phosphorus loss. The trend results did not appear to qualitatively differ across different categories of field size. The results from examining the trends across different categories of median soil wetness index class did show some differences. In the lowest quartile (driest sites), there is a strong indication of downward trends and the slopes were relatively large in percentage terms (-3.3 percent per year). These fields might be expected to be smaller contributors to phosphorus losses because they do not often become saturated and generate runoff. At the other end of the spectrum, the fields with median soil wetness index classes in the highest quartile also had downward trends, but to a lesser degree than the drier class and the statistical significance level was marginal (0.06). In general, these results indicate that efforts to draw down soil phosphorus are modestly effective although future efforts might benefit from a stronger emphasis on reducing soil phosphorus in areas that tend to saturate and generate large runoff losses (e.g., soil wetness index classes 7 to 10). These areas are most prone to phosphorus losses and progress in drawing down soil phosphorus levels has been most limited in these areas. This combination of factors (high soil phosphorus on high soil wetness index classes) should drive prioritization for expenditures on BMPs for nutrient management.
Paired Watershed Study and Other Case Studies
Although the overall effectiveness of the WAP has not been subjected to a comprehensive analysis, several studies have evaluated or assessed components of the various programs. Several studies focused on the performance of individual BMPs, including calf housing and stream exclusion fencing; one paired watershed study distilled the integrated impact of BMPs installed in a one-farm watershed; and there have been several modeling studies. Collectively, these studies represent efforts by the Watershed Protection Program and WAC to determine if and how agricultural BMPs are improving water quality. The section below summarizes the studies and their most relevant conclusions.
TABLE 5-2 Results of Regional Kendall Trend Analysis for Soil Phosphorus Data from Cannonsville Reservoir Watershed
|Category of fields||Number of Fields with Two or More Years Sampled||Slope of Trend in %/yr||Two-Sided Significance Level of Trend|
|Other crops||20||-6.8||not significant|
|Lowest quartile of field size||867||-2.2||<0.001|
|Highest quartile of field size||878||-2.7||<0.001|
|Lowest quartile of median soil wetness index class||957||-3.3||<0.001|
|Highest quartile of median soil wetness index class||902||-0.9||0.060|
Bishop et al. (2005) conducted a paired watershed study of two Cannonsville subwatersheds, one a treatment farm that had significant investment in BMPs, and the other a forested control. They evaluated changes in phosphorus loading attributable to implementation of BMPs that included manure management, rotational grazing, stream livestock exclusion, precision feeding, and improved barnyard infrastructure. Stream water monitoring provided data to calculate phosphorus loads from the 160-ha farm watershed for all runoff events during a two-year pretreatment period and a four-year post-treatment period. Statistical control for interannual climatic variability was provided by matched phosphorus loads from the 86-ha forested watershed. Their results show that the BMPs installed at the farm reduced dissolved phosphorus in storm flow by 43 percent (95 percent confidence interval is 36–49 percent) and particulate phosphorus in storm flow by 29 percent (95 percent confidence interval is 15–41 percent) (see Figure 5-7).
In addition, although the BMPs reduced the phosphorus yield from the farm watershed, the yield (load of phosphorus per unit area) of phosphorus from the farm watershed was still substantially higher than the yield from the forested watershed. For total dissolved phosphorus (TDP), the yield of the farm watershed was 12.3 times the yield of the forested watershed before the BMPs were installed and still 7.2 times the forested watershed yield after BMP installation. For particulate phosphorus (PP) the yields went from 7.9 times down to 6.3 (Table 5-3). Thus, even following recommended farming practices, and with practices and technologies in place, the farm watershed delivered much higher loads of phosphorus than did a forested watershed.
Bishop et al. (2005) report that a second round of BMPs was implemented using Precision Feed Management to attempt to address the phosphorus imbalance associated with the import of purchased feed. The
TABLE 5-3 Loads of Total Dissolved Phosphorus (TDP) and Particulate Phosphorus (PP) from the Farm and Non-Farm Watersheds, During the Pre-Treatment (Pre-BMP) and Post-Treatment (Post-BMP) Periods
|Farm Watersheda||Non-Farm Watershed b|
a The farm treatment watershed is 160.7 Ha in size.
b The non-farm control watershed is 86 Ha in size.
c The pre-treatment (pre-BMP) period extended from June 1993 to May 1995.
d The post-treatment (post-BMP) period extended from November 1996 to October 2000.
e Winter: December 16-April 13; Spring: April 14-June 15; Summer: June 16-September 30; Fall: October 1-December 15; Full-year: January 1-December 31.
Committee is aware that at least six years of data have been collected beyond the data used in the Bishop et al. (2005) paper. Given that Precision Feed Management is an important BMP for the control of phosphorus in farming areas, WAC and NYC DEP would benefit from attempting to evaluate those data along with the original data used in the Bishop et al. (2005) study, as well as expanding the analysis to include the base flow sample data that have been collected throughout the paired watershed study.
The analysis by Bishop et al. (2005) indicate two important findings: (1) that BMPs, if installed and maintained correctly, can produce sustained reductions in farm-level phosphorus yield and (2) it is unlikely that BMPs will reduce phosphorus yield to a level similar to forested areas without reducing the overall phosphorus mass balance. The results of this study, as described in Table 5-3, indicate that even when practices installed are highly effective (as this study suggests they are), there can be a higher yield in the period after the practice was installed simply because the amount and intensity of rainfall might be much greater in the period after installation than the period before. This study is an example of the profound effect that natural variability can have
on the apparent effectiveness of the BMPs. This problem is overcome in this study because of the with- and without- and pre- and post-BMP study design that was utilized.
The Committee was impressed with the very effective paired watershed, pre- and post-implementation design of this study and the use of analysis of covariance to make the study powerful enough to discern the treatment effect in the light of strong interannual variability. The NYC DEP should consider funding more studies of this general type to evaluate the effectiveness of any of their major pollution-reduction strategies. One caveat the Committee would make about this particular study is that it focused only on “event” (storm) results. The Bishop et al. (2005) paper notes that for TDP the event loads constitute about 57 percent of the total annual load; the remaining 43 percent of loads occur under base flow conditions, when streamflow is dominated by soil water and groundwater contributions. Because these base flow loads can be important, especially for the more biologically active dissolved fraction of the phosphorus load, future studies should consider total loads and not just event loads.
In the same farm watershed studied by Bishop, Easton et al. (2008a) largely corroborated these findings using a combination of distributed, mechanistic modeling and long-term monitoring. Their results showed that BMPs reduced dissolved phosphorus loads by 35 percent, similar to Bishop et al. (2005). Perhaps more important, the combined modeling and monitoring approach was able to estimate that the most effective BMPs were those that disassociated pollutant loading areas from areas prone to generating runoff, that is, hydrologically sensitive areas. This dissociation was accomplished primarily via two BMPs: the Nutrient Management Plans, which redistributed manure from hydrologically sensitive areas of the farm to lower-runoff-risk locations, and by altering hydrologic flowpaths (by intercepting shallow interflow with ditching). Attempts to reduce phosphorus content in manure were somewhat less effective, although over the long term, reducing the phosphorus content of manure would help bring the phosphorus mass balance back into alignment.
Hoang et al. (2017) describes the application of the SWAT-Hillslope (SWAT-HS) model to the Cannonsville Reservoir watershed. Farm acreage in the Cannonsville watershed constitutes 19 percent of the land area, while generating, according to the model results, 47 percent of the dissolved phosphorus load and 79 percent of the particulate phosphorus load. To evaluate the impact of BMPs, the authors simulated the effect of two BMPs. One was the use of Nutrient Management Plans to reduce fertilizer and manure applied to croplands, and the other was prescribed grazing to better distribute manure deposited on the landscape by cattle. Using the SWAT-HS model, the authors estimated that these two BMPs should result in a 21 percent reduction in the dissolved phosphorus load and a 35 percent reduction in the particulate phosphorus load, although no data were presented in the paper to corroborate these model simulations. In the future, model-based studies would be strengthened by the collection and analysis of field data to corroborate model results.
Collick et al. (2006) evaluated two dairy calf housing types. The first was standard calf management, where the calves remain with the herd. The second was a recommended BMP, solar calf housing, where the calves are isolated from the herd in an attempt to reduce the pathogen burden of the animals, which is intended to control Cryptosporidium parvum in young calves (the primary source of C. parvum in dairy herds). Although solar calf housing was ineffective in reducing the C. parvum burden in calves, the BMP allowed contaminated calf manure to be isolated from the main barn manure and potentially managed differently and is a way to decrease the number of viable oocysts entering the environment during field spreading. Indeed, Jenkins et al. (1999)
provide evidence that storing calf manure prior to field application (particularly through freeze–thaw cycles) can reduce pathogen survival.
James et al. (2007) evaluated the magnitude of direct in-stream fecal deposits by dairy herds in the Cannonsville watershed and estimated that 2,800 kg of phosphorus are deposited directly in streams, with another 5,600 kg of phosphorus deposited within 10 meters of watershed streams (note that approximately 78,000 kg/yr of phosphorus enter the Cannonsville Reservoir—see Chapter 4). While they did not directly evaluate the impact of excluding livestock from streams (a common BMP), their results point toward the importance of removing livestock from waterbodies. They also estimated that the CREP program (essentially removing livestock from the stream/near-stream area) resulted in a 32 percent reduction in fecal phosphorus deposits, but caution that CREP could be even more effective if herd size, pasturing time, and cattle movement were used to prioritize sites for livestock exclusion.
A New Approach to Manure Management
Observed trends in the trophic status of Cannonsville Reservoir and TP fluxes in the West Branch of the Delaware River upstream of the reservoir (see Chapter 4) suggest that the current strategies for agricultural nonpoint source control of phosphorus are not yet leading to clear improvements in the trophic status of the reservoir. Manure constitutes the major water quality threat, representing 60-85 percent of the land-applied phosphorus in the watershed (Cerosaletti et al., 2004). Approximately 80 percent of the WOH farms have no capacity to store manure; thus, most farms are forced to spread manure on a daily basis. Nutrient Management Plans attempt to direct risky spreading (such as calf manure, spreading in winter, or during times when soils are saturated) to so-called low-risk application areas. Yet, as the NYC watershed region gets wetter due to climate change, a larger percentage of the landscape will begin contributing runoff and pollutant losses. Legacy nutrients that have historically been sequestered in areas of the watershed less prone to loss may become mobilized.
Given these growing risks from manure applications, the Committee suggests a new approach to manure management. The proposed approach is not a regulatory approach, but rather a public-private partnership to collect and treat animal manure and create energy and nutrients that can be marketed to defray some of the cost of this effort. The premise of the approach is that agricultural manure (primarily from dairy operations) should be viewed as a resource to be utilized, rather than a pollutant that needs to be disposed of. It is a resource because of the nitrogen, phosphorus, and carbon that the manure contains, and because manure can be used to produce energy (as methane and/or heat), which can be monetized or used on-farm. Several conversion technologies exist for processing manure in ways that greatly reduce its volume while producing useable co-products (e.g., energy, fertilizer, soil amendment).
There are two primary processes for manure utilization—the use of heat and the use of microbes—and several techniques exist for each process (Szogi et al., 2015). These processes can produce products, such as biochar (see Lehmann et al., 2006; Woolf et al., 2010) and soil amendments, or heat and energy, so called “manure-to-energy” systems that produce primarily methane via anaerobic digestion or thermochemical processes. Some methods are better suited to different types of manure (e.g., liquid content of manure) or scales of operation. The products are often rich in nutrients, so their use and handling is an important consideration for manure utilization operations. If such operations intend to support water quality goals, they must ensure that the products are used in accordance with a Nutrient Management Plan or processed and sold as commercial fertilizer where nutrients are needed.
There are four heat-based processes for using manure to produce energy: combustion, gasification, pyrolysis, and torrefaction, each of which tends to be more appropriate for manure with a low moisture content,
such as poultry litter, because the effort to reduce large liquid contents in the manure is avoided. Microbial processes for manure utilization are via anaerobic digestion. This process is managed in a digester where air is excluded and microbes degrade organic matter in manure into methane (CH4), carbon dioxide (CO2), nitrogen (N2, N2O), and hydrogen sulfide (H2S). Because anaerobic bacteria require oxygen-free environments they are ideally suited for high-liquid manure such as from cows and swine (Tafdrup, 1995).
Each process also has challenges. Thermochemical processes that operate in the presence of oxygen, such as combustion, generate air emissions (e.g., oxides of nitrogen, N2O), while gasification, pyrolysis, and torrefaction produce co-products that are potentially nutrient rich and so must be managed appropriately (Cantrell et al., 2007). Indeed, the greatest challenge in anaerobic digestion is that almost all nutrients from the manure remain in the co-product, although often in more recalcitrant forms (Appels et al., 2011).
If such manure utilization facilities existed in a distributed manner across the watershed (potentially located in several hamlets), the distance that the manure would have to be transported from the farm to the treatment facility could be reduced and any energy produced could be used locally in the hamlet. Manure-to-energy systems are common in Northern Europe. For example, by 2012, 10 percent of livestock manure in the Netherlands was used for energy production (Agro Business Park A/S, 2012) and as of 2017 the Netherlands had 252 digesters with an installed electrical capacity of 219 MW (World Biogas Association, 2017).
The Committee proposes that NYC DEP create a system to incentivize private-sector innovation for community manure treatment in the watershed, with appropriate regulatory oversight. The economics of these technologies are generally considered marginal at present, so investment in such systems needs public support (e.g., NYC DEP). The total costs of such systems includes acquisition of the land for the conversion facility, purchase and installation of the conversion technology, sufficient on-farm storage to accumulate the manure between manure pick-ups, costs to transport the manure, and system operating costs. With present technology, such systems would require a subsidy and a guaranteed market for the outputs, which NYC DEP could help to provide.
NYC DEP has spent over $700,000,000 in capital costs (from 1993 to 2019) to provide advanced wastewater treatment (Table 1-1), reducing nutrient and microbial discharge to extremely low levels. NYC DEP’s annual operating costs for the new, private wastewater treatment plants (WWTPs) are currently around $14,000,000 per year. Given the success of the wastewater treatment effort in the watershed, attention should now be turned to the wastes produced from commercial animal production. Based on an estimated 17,402 animal units in the NYC watershed in 2017 (Graf, 2018) and the contribution of about 11 kg phosphorus per animal unit per year (USDA NRCS, 1995), about 191,000 kg/yr of phosphorus is excreted by the animals in the watershed. For comparison, the total output of phosphorus from all of the WWTPs in the WOH watersheds in 1994 (prior to the upgrading to tertiary treatment) was about 7,030 kg/yr (see Table 8-4). This means that the total output of phosphorus from livestock (virtually all are WOH) was about 27 times the discharge of phosphorus from humans (by way of the WOH WWTPs) prior to the WWTP upgrades. The current output of phosphorus from the upgraded WWTPs is about 162 kg/yr (Table 8-4). Even though it is likely that a substantial portion of the phosphorus from livestock is removed by the growth and harvest of row crops and pasture, or stored in the soil and groundwater, the contribution of phosphorus from livestock is far greater than the phosphorus from WWTPs. Note that in both cases (WWTP upgrades and potential conversion of animal waste) the benefits are not limited to phosphorus removal, but also provide reductions in nitrogen, carbon, pathogens, and biological oxygen demand.
A public-private partnership to collect and treat animal wastes should be created and would bring substantial benefits to the NYC water supply and reduce the likelihood that a costly filtration system would be required in the future. The use of NYC DEP funds to support the collection and treatment of animal wastes (as is already the case with human wastes) is a worthwhile investment in support of watershed protection and would also contribute to community vitality through the jobs created.
Currently, BMP prioritization is based primarily on the order that the pollutant categories were established and on farm rank (see Whole Farm Plan section). WAC currently does not have data on which BMPs present the best economic or water quality value to the program.
Given the significant investment NYC DEP has made in BMPs via the WAP, they should move toward performance-based metrics to characterize BMPs. Conventional financial assistance (cost-share) programs compensate people for the direct financial costs to install approved BMPs. The pollutant removal effectiveness is unknown by those implementing the practice and performance is assumed if operated and maintained in compliance with recommendations. In contrast, pay-for-performance systems provide financial payments based on the outcome achieved. In the case of water quality, performance can be defined as reductions in effluent pollutant loads ($/kg of reduction). Paying for outcomes, rather than practices, offers the potential to enhance pollutant removal effectiveness and reduce pollutant control costs while supporting agricultural production (Shabman et al., 2013; Shortle, 2017). Several studies have shown that, in general, paying for performance (e.g., amount of nutrient loss reduced) is more cost-effective than basing payments on practice costs (Ribaudo et al., 1999; Savage and Ribaudo, 2016). Producers who can provide the greatest reductions at the lowest cost have the largest economic incentive to act. This means that farmers who may not have traditionally participated in conservation programs might have a strong incentive to do so. Such an approach is “self- selecting” in that those who can provide the most environmental benefit at least cost stand to gain the greatest economic benefit. In addition, performance-based payments could provide more flexibility in how a particular environmental service is produced. Practice-based payments tend to limit choice to practices that are cost-shared, while performance-based policies award innovations that lower cost.
One challenge to pay-for-performance is quantifying performance at relevant spatial and temporal scales to support payment for nutrient reductions. For instance, it is unreasonable to expect quantification of nutrient and sediment reduction at the field scale (this would require extensive monitoring), but perhaps some prioritized reach-level quantification of the aggregate impact of BMPs is feasible. Field-level measurement or modeling tools for estimating BMP performance are needed with performance-based policies.
A Role for Watershed Modeling
Within the existing NYC DEP watershed modeling framework, two tools are utilized to predict water quality and how it is expected to change under various land management and BMP activities: VSLF (also referred to as GWLF-VSA, or GWLF, Schneiderman et al., 2007) and SWAT-HS (Hoang et al., 2017, 2019). Both tools may be used in some manner to predict BMP performance; however, both are constrained by shortcomings. Employing VSLF to understand BMP performance would be unlikely to yield significant insight into processes controlling nutrient or sediment loss because it does not utilize process/mechanistic nutrient or sediment routines to predict pollutant loss; it uses a simple mean expected concentration approach for each land use. That is, each land use is described by a characteristic nutrient concentration (also referred to as export coefficient), derived from literature or from empirical data. For example, phosphorus concentrations that VSLF uses for agricultural runoff range from 0.20 mg P/L for pasture to 0.26 mg P/L for corn, to 5.10 mg P/L for barnyards (Schneiderman et al., 2002); these values may or may not vary by season and are generally calibrated from the initial values. Thus, it is at best a calibration artifact to assume that prediction of nutrient loss from any given parcel reflects reality. Using VSLF to evaluate management-type BMPs, such as no-till, nutrient management planning, exclusionary fencing, or pasture management would provide less valuable information because these BMPs reduce the pollutant at the source. That is, they change (presumably reduce) the amount of a pollutant available for transport from a field, and thus require a change in the mean expected concentration in runoff to model.
However, VSLF could be used to assess the performance of so-called structural BMPs that affect the hydrologic response. For instance, BMPs that alter flow paths, such as ditching, diversions, or drainage (as in Easton et al., 2008a) affect nutrient loss by altering the hydrologic response (reducing runoff perhaps), and thus
reducing the flux of a pollutant leaving a field. This approach assumes no change to the mean expected concentration from a field, only a change in the hydrologic response.
SWAT-HS improves upon SWAT by introducing an aquifer that allows lateral transfer of subsurface water across the landscape based on the variable source area (VSA) concept. This modification better accounts for water movement between modeling units, thus improving the prediction of where saturated, runoff generating areas occur. SWAT-HS could be leveraged to provide useful information on BMP performance because it incorporates mechanistic-based nutrient routines. In addition to being able to assess structural BMPs that affect the hydrological response, similar to VSLF, it would add the ability to potentially capture changes to nutrient concentrations in the soil that result from structural BMPs, for instance, drainage, which in addition to reducing surface runoff increases nitrification in the soil. For management-type BMPs (e.g., nutrient management, no till, rotational grazing), SWAT-HS offers clear advantages over VSLF. SWAT can mechanistically simulate nutrient process occurring in the soil and plant or crop, it can model the buildup or draw-down of soil nutrients, and it allows the user to explicitly detail the amount of manure or fertilizer applied to each hydrologic response unit (HRU) or the type of tillage used. Thus, BMPs such as Nutrient Management Plans, which reduce the amount of nutrient in the system, or no-till, which increases soil carbon and structure, can be evaluated.
Ultimately, both models are constrained by issues of scale. Both employ some derivative of the HRU watershed discretization scheme, where an HRU is an overlay of unique soil type, terrain class (soil wetness index class), and land use. Thus, HRUs with the same soil type, terrain class, and land use within the same subbasin are assumed to have identical properties and thus identical parameters in the model. As a result, BMPs cannot be explicitly modeled on an individual field but can be evaluated across morphometrically similar units (e.g., one can assess Nutrient Management Plans across an individual HRU, which may encompass many individual fields). There are options to potentially increase the spatial resolution of SWAT-HS by reducing the threshold size to define a subbasin, which would reduce the number of unique HRUs in any given subbasin. This change increases the computational costs of model runs and gathering information to parameterize the model at this scale could prove challenging. However, it is the Committee’s opinion that SWAT-HS, when properly parametrized and applied, offers clear benefits over more simplistic models such as VSLF/GWLF, in particular with respect to BMP assessment.
Modeling to Enable Prioritization of BMPs
BMP implementation programs typically work based on factors including the willingness of landowners to participate and distribution of financial incentives. However, implementing BMPs on areas of the landscape that produce the most nonpoint source pollution is essential to achieve pollutant reduction goals.
Identification and prioritization of lands for BMP implementation can be based on the pollutants. If phosphorus and sediment are the concern, BMPs that focus predominantly surface or overland flow pathways are preferred. Nitrogen moves by both surface and subsurface paths, with subsurface flow being more difficult to treat, such that source reductions are better BMP choices for nitrogen.
Alternatively, prioritizing lands for BMP implementation can be based on the source area. For surface pathway pollutants, flow-path or terrain models (that incorporate landscape connectivity, hydrologic distance, and soil depth) can identify hydrologically sensitive areas, and when intersected with land use can provide an estimate of where critical pollutant source areas occur. For groundwater pathway pollutants, identifying where recharge areas occur (perhaps from soil drainage class and restricting capacity) and intersecting these with land use data can provide an estimate of where critical pollutant source areas occur.
Several tools in the NYC DEP modeling framework could be used to prioritize areas of the watershed for BMP application. Both the VSLF (GWLF) and SWAT-HS models (Hoang et al., 2017) are underpinned by a topographic index terrain model (Beven and Kirkby, 1979). Specifically, they use a soil topographic index (TI):
where α is the upslope contributing area per unit of contour line (meter), tanβ is the topographic slope of the landscape, and T is the transmissivity (soil depth × saturated soil hydraulic conductivity, Ks) of the uppermost layer of soil (m2/d). Shallower soils and those with lower Ks values will have lower soil water storage capacities and thus higher runoff probabilities. This approach accounts for hydrological disconnection of overland flow due to re-infiltration where soils are unsaturated. Topographic indices have been found to improve predictions of soil moisture, runoff production, and pollutant delivery from diffuse sources compared to approaches that do not consider topography, such as waterbody proximity (Buchanan et al., 2014; Hahn et al., 2014; Easton et al., 2007a,b; 2008b; 2011; Schneiderman et al., 2007). Lyon et al. (2004) found that topographic indices described the evolution of shallow water tables in the Catskill Mountains of New York and that this shallow water table was the primary control on Variable Source Area formation. Larger upslope drainage areas and shallower slopes will produce larger TI values indicating higher runoff probability (Quinn et al., 1991).
Since both models employ overlays of soil type, topographic index, and land use, the dominant controls on critical source area formation are embodied in model predictions. Thus, areas suitable for BMPs (and perhaps acquisition depending on the size of a parcel that is eligible) could be prioritized without new model development.
Prioritizing via modeling would be more of a challenge. As discussed previously, VSLF does not utilize process/mechanistic nutrient or sediment routines to predict pollutant loss; it uses a simple mean expected concentration approach for each land use. SWAT-HS could likely be leveraged to provide estimates of areas suitable for BMPs, because it incorporates process- or mechanistic-based nutrient routines, although calibrating the model at a fine enough resolution to provide useful information could prove challenging.
NYC DEP staff, or researchers funded by NYC DEP, would benefit from exploring the ability of SWAT-HS to simulate the effects of BMPs by doing studies that compare observed loadings (by storm event and/or by month) at long-term stream monitoring sites with loadings estimated by SWAT-HS using the actual history of BMP implementation in the monitored watershed. (See more discussion of the importance of verification of the SWAT-HS model in Chapter 12). The key question to be answered by such studies is whether the SWAT-HS can even roughly capture trends in pollutant outputs that occur over a period of years. Watersheds, such as those used in the Bishop et al. (2005) study, may provide an ideal setting for such analyses. Such studies can only be accomplished when the following conditions are met:
- Periods of monitoring include both before and after BMP implementation, with BMPs implemented for at least several years in order to capture a reasonable range of weather conditions.
- The use of a “treatment” watershed (where BMPs are widely implemented) and a “control” watershed (where they are not implemented) is crucial. The shorter the distance between the treatment and control watersheds, the better the chance of obtaining meaningful results.
- The extent of BMP implementation in the treatment watershed should be large (probably greater than 20 percent). Even a highly effective BMP applied to 5 percent of the monitored watershed is unlikely to produce meaningful evidence about the utility of the BMP being evaluated.
- The evaluation should be based on loadings and not on concentrations, because loadings are what influences conditions in downstream reservoirs.
These study conditions are not easily met but are necessary. NYC DEP could seek to collaborate with others and identify multiple funding sources for such studies. The results of such studies will be very beneficial within the NYC watershed and well beyond it. The combination of modeling with analysis of small-watershed empirical results is what is needed to provide reliable tools for prioritization of effective watershed protection strategies.
Besides source pollutant reductions, BMP implementation stands at the center of NYC DEP’s efforts to meet SWTR requirements. The Committee’s examination in Chapter 4 of the recent history of reservoir phosphorus concentrations in Cannonsville shows that there is little progress being made in reducing these
phosphorus concentrations despite large expenditures being made on BMP implementation upstream. This suggests that BMPs may not be performing as assumed. One possible reason is uncertainty regarding BMP efficiencies. That is, efficiency estimates have multiple embedded assumptions about hydrologic setting, flow rates, maintenance conditions, among others, that may not be reflective of on-the-ground conditions. Second, since BMP implementation is voluntary, WAC cannot effectively prioritize BMP installation in localized source areas. Researchers have noted that areas of high nutrient loss are site specific and highly localized. If BMPs are applied in lower-risk areas (because participation is voluntary) rather than in areas supplying greater nutrient loads, BMP effectiveness may be overestimated. Third, the program does not account for the lag time between implementation and pollutant reductions. Given groundwater and intermittent surface runoff, lag times may be longer than expected.
Characterizing BMP performance uncertainty requires systematic identification of the factors that make outcomes variable and uncertain. BMP pollutant removal performance typically consists of multiple components, including the variability of incoming loads, site conditions, installation and maintenance, weather, types of removal processes (e.g., storage, biological transformation) and multiple loss pathways (e.g., surface runoff, groundwater leaching, atmospheric losses). The ability to observe and accurately measure performance differs across BMPs. BMPs also vary widely in the complexity of the various factors involved in pollutant removal. WAC working with NYC DEP could identify some basic estimates on the probability distribution of BMP performance effectiveness (e.g., first and third quartiles or one standard deviation from the mean estimate, and an assessment of the shape of the distribution or skewness). This would be sufficient to allow the NYC DEP modeling group to represent a range of outcomes from any given BMP implementation (e.g., probability of meeting load reduction goals). With information on the distribution of BMP performance estimates, risk management of BMP investment decisions becomes possible. With information on the probabilities of obtaining specific levels of performance, decision-makers can evaluate BMP costs against the risks of not achieving load reduction goals. Ultimately, this can better inform decisions about allocation of resources to achieve water quality goals.
Another means to incorporate BMP performance uncertainty would be to define site and operational characteristics and resultant BMP performance. Ideally these characteristics would be determined from on-the-ground BMPs that WAC or NYC DEP is monitoring (see recommendations). This could be used in a modeling framework to assess BMP performance that may vary across different soil, topography, climate, etc., of the basin. If models are the primary means to evaluate BMP performance (and uncertainty), the program should leverage process-based models to evaluate BMP performance under a range of conditions. This modeling would ideally occur at a relatively fine scale (e.g., farm- or field-scale) in order to fully decompose BMP performance. An important element in determining the adequacy of BMP efficiency estimates is through comparison of observed nutrient trends using an appropriate statistical model versus hindcasted trends generated by the models, in which the models use the actual history of BMP implementation along with the actual history of precipitation and temperature.
Agricultural Trends in the Watershed
The past decade in the WOH region has seen a shift in the types and footprints of farms, with many dairy operations transitioning to predominantly grass-fed beef operations. The transition from dairy to beef presents both challenges and opportunities for the WAP. Dairy cows tend to be concentrated in or near the barnyard due to the frequency of milking, making their manure more concentrated in that area, a potential water quality problem. Yet with correct management, this manure can be easier to clean up and store for spreading, or for collecting for input to a manure-conversion process. Beef cattle tend to be pastured for more of the year, which results in more diffuse manure deposition, potentially in areas of the farm where soil phosphorus levels are lower; this can reduce the risk of phosphorus loss via runoff. However, the more diffuse deposition also presents challenges, primarily because cattle spend most of their time in streams or near-stream areas (James et al., 2007); in areas without exclusionary fencing, beef cattle can increase water quality problems. Perhaps
more concerning is the impact of the 19 percent increase in animal units on the phosphorus mass balance in the watershed over the five-year period covered by the USDA Census of Agriculture (USDA-NASS, 2017). While lactating dairy cows produce substantially more manure nitrogen than beef cattle (0.45 lbs/day per 1,000-lb dairy animal unit vs 0.21 lbs/day per 1,000-lb beef animal unit), beef cattle tend to produce more manure phosphorus than dairy cows (0.11 lbs/day per 1,000-lb beef animal unit vs 0.07 lbs/day per 1,000-lb dairy animal unit), an increase of 36 percent (USDA-NRCS, 1992). Thus, despite the 4 percent decrease in total farm acres operated in the watershed (USDA NASS, 2017), the increase in beef livestock constitutes a considerable potential increase in the phosphorus mass balance on the remaining acres in agricultural use, particularly in the Cannonsville watershed.
Nutrient Mass Balance and Legacy Nutrients
Large mass balance and legacy nutrient issues will exist where inputs of feed and fertilizer exceed the assimilative capacity of the landscape, as they do in the Cannonsville basin, as demonstrated by the existence of soils with phosphorus levels well above agronomic needs. Continued growth in intensive animal agriculture will exacerbate this issue. Several issues result from this mass balance problem. First, the loads from these areas are not adequately reflected in the NYC DEP modeling approach and conventional BMPs are unlikely to overcome the mass balance issue, as very few are designed to bring the mass balance back into alignment. Second, nutrient loads from manure are not adequately reflected in analyses completed by WAP (except for the Precision Feed Management program) or in the NYC DEP modeling tools (see Hoang et al., 2019). The Committee also suspects that large stores of nitrogen and phosphorus in soils and groundwater exist throughout the watershed given the history of livestock-intensive agriculture. This suspicion is confirmed for phosphorus by the proportion of farm fields with high or very high soil phosphorus concentrations (see Figure 5-2). Under the most optimistic assumptions, the drawdown in nutrients, particularly phosphorus levels in soils and groundwater, could take decades.
Failing to adequately account for the impact of these stores of nutrients across the landscape may misrepresent program effectiveness and should be addressed via several synergistic activities. These activities include manure-to-energy or manure byproduct utilization with applications of the byproduct within (or outside) the watershed based on agronomic needs (discussed previously). Second, increased manure storage would substantially improve water quality by preventing manure application to frozen or saturated soils or at times of high risk, even though increased storage would not alter the overall nutrient mass balance, except for some chemical species conversions. Finally, BMPs that can treat/reduce the impact of legacy nutrients—such as wetlands (Kellogg et al., 2005), woodchip bioreactors (Easton et al., 2019), or the use of slag from ore smelting (Kostura et al., 2005)—need to be developed and used.
Climate change has potential to impact hydrology and diffuse nutrient export from agricultural landscapes (Ahmadi et al., 2014; Howarth et al., 2006). In the humid temperate eastern United States, climate predictions suggest that precipitation quantity (during the winter/spring), and intensity (during the growing season) will continue to increase, which may increase diffuse nutrient and sediment export from agricultural landscapes (Chang et al., 2001; Cousino et al., 2015; Wagena and Easton, 2018). This increased export could have several harmful consequences for reservoirs: accelerated eutrophication and harmful algal blooms (Burgin and Hamilton, 2007) and undesirable changes in the river structure and function. In addition, the loss of valuable nutrients and topsoil from agricultural fields decreases productivity or increases management intensity (Lal, 1998).
Changes in precipitation and temperature alter the timing and magnitude of runoff, soil moisture, and biogeochemical cycles (Gleick, 1989). For instance, common processes in the nutrient cycling are to a large
extent controlled by factors that climate change influences, such as soil temperature and soil moisture (Butterbach-Bahl and Dannenmann, 2011). Similarly, increased soil temperatures and moisture content can influence the sorption and desorption of phosphorus, as well as immobilization and mineralization rates, all factors affecting phosphorus export (Sheppard and Racz, 1984). It is also well established that sediment transport is affected by soil moisture (Wiggs et al., 2004) and by precipitation amount and intensity (Römkens et al., 2002), all of which are expected to change in the region. These changes to precipitation and temperature are likely to alter the timing and magnitude of streamflow and nutrient/sediment production and transport in the watershed. For instance, increased spring nutrient export from the watershed and delivery to reservoirs can set up conditions that cause particularly acute summer hypoxia and drier conditions in the summer and fall have been shown to increase the buildup of soil nutrients that can subsequently be flushed from the system when wet conditions return (Wetz and Yoskowitz, 2013). Temperature changes can alter nutrient cycling, plant growth, evapotranspiration, and soil water content, which all affect the availability and transport of nutrients from agricultural fields.
BMPs designed and installed to handle historic weather conditions may not function under a changing climate. Of particular interest is the impact of climate change on hydrologically active areas of the landscape that contribute disproportionally to watershed nutrient export (e.g., critical source areas, CSAs), where active hydrologic transport and high nutrient availability coincide (Groffman et al., 2009). In the WOH system, Pradhanang et al. (2013) found that earlier snowmelt and reduced snowpack advance the timing and increase the magnitude of discharge in the winter and early spring (November-March) and greatly decrease monthly streamflow later in the year, resulting in increased flashiness during winter and lower flows in the summer, both of which negatively impact water quality.
While NYC DEP recognizes the potential consequences of climate change (see NYC DEP, 2009), their focus is almost entirely on infrastructure and operational impacts, with very little focus on how climate change may affect agriculture, agricultural BMPs, and the resultant changes to water quality. Clearly the scientific literature points toward potentially large changes in how climate affects both agricultural production and water quality, which needs to be explicitly considered in the decision-making process of the WAP.
WAC has done a commendable job of bringing watershed producers into the WAP, and has had continued success in engaging relevant stakeholders. These successes should be recognized and further leveraged to increase program effectiveness. Increasing the collaboration between WAC and NYC DEP will result in more efficient allocation of resources, improved communication, and ultimately improved water quality. Several recommendations are made to improve the WAP suite of programs, some of which are relatively straightforward and easily implemented, and some of which will require more strategic and sustained effort and a reallocation of current resources or increased funding (as discussed further in Chapter 14).
The Committee’s analysis of soil phosphorus data collected from agricultural lands in the Cannonsville watershed shows that concentrations have fallen slowly over the last decade. These data were collected by WAC at great expense and are certainly useful for providing guidance to individual farmers. However, they also are useful to NYC DEP and WAC regarding the overall progress in reducing the legacy phosphorus in agricultural soils. As discussed in Chapter 12, the analysis of the large data sets collected as a part of the Watershed Protection Program could become a tremendous asset, but only when subjected to continuous summary and analysis. NYC DEP and WAC should regularly analyze soil phosphorus and other similar data on a continual basis to guide ongoing improvement in phosphorus management efforts.
The Watershed Agricultural Council (WAC) should objectively identify and rank areas of the landscape for BMP installation, prioritizing areas expected to produce the largest runoff and pollutant loadings. The program could use modeling, following the recommendations on how modeling can be used to inform BMP design and implementation (in this chapter) and/or monitoring, following the recommendation in
Chapter 12. One easily implemented area where BMP prioritization may have a large impact is in the Nutrient Management Program; WAC should focus on those fields that have both high soil phosphorus concentrations and large runoff losses (see Figure 5-5).
To combat phosphorus mass balance problems, the Watershed Agricultural Council should develop a public-private partnership to turn manure into energy and/or other useful byproducts. New businesses could collect all available farm manure for significant resource recovery (e.g., phosphorus, nitrogen, carbon, and energy), to be utilized inside or outside of the watershed as appropriate. An important incentive for the private sector to participate is for NYC DEP to provide a guaranteed market (at a known price) for the products created through such systems. Providing subsidies to manure collection and processing, along with priority cost sharing for exclusionary fencing of riparian zone and watercourses, will be vital tools to achieve the goals of reducing phosphorus as well as nitrogen, carbon, and pathogens in a manner that is consistent with enhancing the economic vitality of farming in the watershed.
While the manure utilization program (Recommendation 3) is being developed and implemented, the Precision Feed Management Program should rapidly expand to include a much higher percentage of farms and animal units, which would reduce the mass of manure-derived nutrients in the watershed.
The Watershed Agricultural Council (WAC) should replace the current, 11-level BMP prioritization scheme, which is unduly complicated, with something much simpler and easier to implement, based on the priorities identified in this report. The first priority is manure management; WAC should prioritize removal of livestock from streams and near-stream areas to prevent direct delivery of manure to streams. WAC, working closely with NYC DEP, should make a concerted effort to collect and store manure from confined animals so that it can either be transported and converted to energy or spread at times and in areas of low risk of loss. The second priority is nutrient management, both chemical and manure. Nutrients should be applied in the right amount, in the right place, at the right time, and in the right manner. The third and fourth priorities are sediment management and pesticide management (including storage, handling, and application).
Transitioning program metrics from measures of output (e.g., number of Nutrient Management Plans written, or farms enrolled) to more outcome-oriented metrics (such as estimated nutrient loads prevented from entering streams) that have direct relevance to water quality protection should occur across all WAP programs. One way this could be approached is to incentivize participants in WAP to produce more-effective and lower-cost outcomes by requiring pay-for-performance monitoring. Pay-for-performance incentivizes stakeholders to demonstrate nutrient and sediment reductions beyond what is typically assumed during standard BMP program implementation (e.g., effectiveness of BMPs are measured instead of assumed, similar to metrics of Bishop et al., 2005).
A new and different approach to assessing BMP performance is needed within the Watershed Agricultural Program. First, the program should perform or incentivize site-specific monitoring of pre- and post-BMP performance. This includes producing studies that take the Bishop et al. (2005) report and adds analysis of the later Precision Feed Management efforts and also includes undertaking one or more additional paired watershed studies to add to the empirical evaluation of common BMPs in the watershed. Second, the program should also perform more robust analysis and statistical estimation of BMP performance (see Chapter 12 for more information). Ultimately, the program should transition from compliance monitoring to performance monitoring of BMP effectiveness, again using Bishop et al. (2005) as a template.
The Watershed Agricultural Council (WAC) and the New York City Department of Environmental Protection (NYC DEP) should jointly develop a climate action plan for agriculture that clearly defines potential impacts, proposes actions to mitigate those impacts, and devises an adaptation strategy that ensures
that agriculture does not contribute disproportionally to water quality degradation. An approach that addresses climate change could include accounting for the expected increase in precipitation intensity and temperature, overland flow, and pollutant transport expected from climate change in future BMP designs to ensure their correct functioning under changing conditions. WAC and NYC DEP should also evaluate the potential impact on agricultural production, such as shifts in crop range and growing season length, among others. This is best accomplished with a modeling program capable of representing the complex interactions between climate and agricultural production systems.
Agro Business Park A/S. 2012. Livestock manure to energy: Status, technologies and innovation in Denmark. https://www.agropark.dk 〉admin 〉public 〉download 〉Livestock-Manur…
Ahmadi, M., R. Records, and M. Arabi. 2014. Impact of climate change on diffuse pollutant fluxes at the watershed scale. Hydrological Processes 28(4):1962-1972.
Appels, L., J. Lauwers, J. Degrève, L. Helsen, B. Lievens, K. Willems, J. Van Impe, and R. Dewil. 2011. Anaerobic digestion in global bio-energy production: Potential and research challenges. Renewable and Sustainable Energy Reviews 15(9):4295-4301.
Beven, K. J., and M. J. Kirkby. 1979. A physically-based, variable contributing area model of basin hydrology. Hydrological Sciences Bulletin 24:43–69.
Bishop, P. L., W. D. Hively, J. R. Stedinger, M. R. Rafferty, J. L. Lojpersberger, and J. A. Bloomfield. 2005. Multivariate analysis of paired watershed data to evaluate agricultural best management practice effects on stream water phosphorus. Journal of Environmental Quality 34:1087–1101.
Buchanan, B. P., Fleming, M., Schneider, R. L., Richards, B. K., Archibald, J., Qiu, Z., Walter, M. T., 2014. Evaluating topographic wetness indices across central New York agricultural landscapes. Hydrology and Earth System Sciences 18:3279–3299.
Burgin A., and S. Hamilton. 2007. Have we overemphasized the role of denitrification in aquatic ecosystems? A review of nitrate removal pathways. Frontiers in Ecology and the Environment 5(2):89-96.
Butterbach-Bahl, K., and M. Dannenmann. 2011. Denitrification and associated soil N2O emissions due to agricultural activities in a changing climate. Current Opinion in Environmental Sustainability 3(5):389-395.
Cantrell, K., K. Ro, D. Mahajan, M. Anjom, P. G. Hunt. 2007. Role of thermochemical conversion in livestock waste-to-energy treatments: Obstacles and opportunities. Industrial and Engineering Chemistry Research 46(26):8918-8927.
Cerosaletti, P., G. Fox, G., and L. Chase. 2004. Phosphorus reduction through precision feeding of dairy cattle. Journal of Dairy Science 87:2314–2323
Cerucci, M., and S. Pacenka. 2003. The application of SWAT and GWLF to small scale agricultural watersheds. In: Watershed Management for Water Supply Systems: Proceedings of the American Water Resources Association 2003 International Water Congress, June 29-July 2, New York, NY. M. J. Pfeffer, D. J. Van Abs, and K. N. Brooks (eds.).
Chang, H., B. M. Evans, and D. R. Easterling. 2001. The effects of climate change on stream flow and nutrient loading. Journal of the American Water Resources Association 37(4):973-985.
Collick, A. S. 2003. Evaluation of solar calf housing for the presence and survival of Cryptosporidium parvum in the NYC watershed. Report to the Watershed Agricultural Program.
Collick, A. S., E. A. Fogarty, P. E. Ziegler, M. T. Walter, D. D. Bowman, and T. S. Steenhuis. 2006. Survival of Cryptosporidium parvum oocysts in calf housing facilities in the New York City watersheds. Journal of Environmental Quality 35:680-687.
Cousino, L. K., R. H. Becker, and K. A. Zmijewski. 2015. Modeling the effects of climate change on water, sediment, and nutrient yields from the Maumee River watershed. Journal of Hydrology: Regional Studies 4(B): 762-775.
Easton, Z. M., P. Gerard-Marchant, M. T. Walter, A. M. Petrovic, and T. S. Steenhuis. 2007a. Identifying dissolved phosphorus source areas and predicting transport from an urban watershed using distributed hydrologic modeling. Water Resources Research 43:W11414.
Easton, Z. M., P. Gérard-Marchant, M. T. Walter, A. M. Petrovic, and T. S. Steenhuis. 2007b. Hydrologic assessment of an urban variable source watershed in the Northeast United States. Water Resources Research 43:W03413.
Easton, Z. M., M. T. Walter and T. S. Steenhuis. 2008a. Combined monitoring and modeling indicate the most effective agricultural best management practices. Journal of Environmental Quality 37:1798–1809.
Easton, Z. M., D. R. Fuka, M. T. Walter, D. M. Cowan, E. M. Schneiderman, and T. S. Steenhuis. 2008b. Re-Conceptualizing the Soil and Water Assessment Tool (SWAT) model to predict runoff from variable source areas. Journal of Hydrology 348:279-291.
Easton, Z. M., M. T. Walter, D. R. Fuka, E. D. White, and T. S. Steenhuis. 2011. A simple concept for calibrating runoff thresholds in quasi-distributed variable source area watershed models. Hydrological Processes 25:3131-3143.
Easton, Z. M., E. M. Bock, and K. Stephenson. 2019. Feasibility of employing woodchip bioreactors to treat legacy nutrients in emergent groundwater. Environment Science and Technology 53(21):12291-12299.
Fiorellino, N., R. Kratochvil, and F. Coale. 2017. Long-term agronomic drawdown of soil phosphorus in mid-Atlantic coastal plain soils. Agronomy Journal 109:455-461.
Gleick, P. H. 1989. Climate change, hydrology, and water resources. Reviews of Geophysics 27(3):329-344.
Graff, J. 2018. Watershed Agricultural program. Presentation at Meeting 2 of the NASEM Committee to Review the NYC DEP Watershed Protection Program. Rhinebeck, NY. October 24-26.
Groffman, P. M., Butterbach-Bahl, K., Fulweiler, R. W., Gold, A. J., Morse, J. L., Stander, E. K., Tague, C., Tonitto, C., and Vidon, P. 2009. Challenges to incorporating spatially and temporally explicit phenomena (hotspots and hot moments) in denitrification models. Biogeochemistry 93(1):49-77.
Hahn, C., V. Prasuhn, C. Stamm, D. G. Milledge, and R. Schulin. 2014. A comparison of three simple approaches to identify critical areas for runoff and dissolved reactive phosphorus losses. Hydrology and Earth System Sciences 18:2975–2991.
Helsel, D. R., and L. M. Frans. 2006. Regional Kendall test for trend. Environmental Science and Technology 40(13):4066-4073.
Hoang, L., E. M. Schneiderman, R. Mukundan, K. E. Moore, E. M. Owens, and T. S. Steenhuis. 2017. Predicting saturation-excess runoff distribution with a lumped hillslope model: SWAT-HS. Hydrological Processes 31(12):2226-2243.
Hoang, L., R. Mukundan, K. E. Moore, E. M. Owens, and T. S. Steenhuis. 2019. Phosphorus reduction in the New York City water supply system: A water quality success story confirmed with data and modeling. Ecological Engineering 135:75-88.
Howarth, R., D. Swaney, E. Boyer, R. Marino, N. Jaworski, and C. Goodale. 2006. The influence of climate on average nitrogen export from large watersheds in the Northeastern United States. Pp. 163-186 In: Nitrogen Cycling in the Americas: Natural and Anthropogenic Influences and Controls. L. Martinelli and R. W. Howarth (eds.). Springer.
James, E., O. Kleinman, T. Veith, R. Stedman, and A. Sharpley. 2007. Phosphorus contributions from pastured dairy cattle to streams of the Cannonsville watershed, New York. Journal of Soil and Water Conservation 62(1):40-47.
Jenkins M., M. Walker, D. Bowman, L. Anthony, and W. Ghiorse. 1999. Use of sentinel system for field measurement of Cryptosporidium parvum oocyst inactivation in soil and animal waste. Applied Environmental Microbiology 565(5):1998-2005.
Kellogg, D. Q., A. J. Gold, P. M. Groffman, K. Addy, M. H. Stolt, and G. Blazejewski. 2005. In situ ground water denitrification in stratified, permeable soils underlying riparian wetlands. Jourmal of Environmental Quality 34:524-533.
Kostura, B., H. Kulveitova, and J. Leško. 2005. Blast furnace slags as sorbents of phosphate from water solutions. Water Research 39(9):1795-1802.
Lal, R. 1998. Soil erosion impact on agronomic productivity and environment quality. Critical Reviews in Plant Sciences 17(4):319-464.
Lehmann, J., J. Gaunt, and M. Rondon. 2006. Bio-char sequestration in terrestrial ecosystems – a review. Mitigation and Adaptation Strategies for Global Change 11:403-427.
Lyon, S. W., M. T. Walter, P. Gerard-Marchant, and T. S. Steenhuis. 2004. Using a topographic index to distribute variable source area runoff predicted with the SCS curve-number equation. Hydrologic Processes 18:2757–2771.
NYC DEP (New York City Department of Environmental Protection). 2006. Watershed Agricultural Program Research Summary and Report.
NYC DEP. 2009. Climate Change Program Assessment and Action Plan.
NYC DEP. 2019. Current Summaries of Watershed Protection Programs. January.
Pradhanang, S., R. Mukundan, E. M. Schneiderman, M. Zion, A. Anandhi, D. C. Pierson, A. Frei, Z. M. Easton, D. R. Fuka, and T. S. Steenhuis. 2013. Streamflow responses to projected climate change in New York City water supply watershed. Journal of American Water Resource Association 49(6):1308‐1326.
Quinn, P., K. Beven, P. Chevallier, and O. Planchon. 1991. The prediction of hillslope flow paths for distributed hydrological modelling using digital terrain models. Hydrological Processes 5:59-79.
Ribaudo, M., R. Horan, and M. E. Smith. 1999. Economics of Water Quality Protection from Nonpoint Sources: Theory and Practice, No 33913, Agricultural Economics Reports. USDA Economic Research Service. https://www.ers.usda.gov/publications/pub-details/?pubid=41066.
Römkens, M. J. M., K. Helming, and S. N. Prasad. 2002. Soil erosion under different rainfall intensities, surface roughness, and soil water regimes. Catena 46(2–3):103-123.
Savage, J., and M. Ribaudo. 2016. Improving the efficiency of voluntary water quality conservation programs. Land Economics 92(1):148-166.
Schneiderman, E. M., D. C. Pierson, D. G. Lounsbury, and M. S. Zion. 2002. Modeling the hydrochemistry of the Cannonsville Watershed with Generalized Watershed Loading Functions (GWLF). Journal of the American Water Resources Association 38(5):1323–1347.
Schneiderman, E. M., T. S. Steenhuis, D. J. Thongs, Z. M. Easton, M. S. Zion, G. F. Mendoza, M. T. Walter, and A. L. Neal. 2007. Incorporating variable source area hydrology into the curve number based Generalized Watershed Loading Function model. Hydrological Processes 21:3420-3430.
Shabman, L., S. Lynch, and E. Boughton. 2013. Acquiring water services from Northern Everglades ranch-lands: Assuring buyers that that they get what they pay for. Rangelands 35(5):88-92.
Sheppard, S., and G. Racz. 1984. Effects of soil temperature on phosphorus extractability. I. Extractions and plant uptake of soil and fertilizer phosphorus. Canadian Journal of Soil Science 64(2):241-254.
Shortle, J. S. 2017. Policy reforms needed for better water quality and lower pollution control costs. Choices 32(4):1-7.
Szogi, A., M. B. Vanotti, and K. S. Ro. 2015. Methods for treatment of animal manures to reduce nutrient pollution prior to soil application. Current Pollution Reports 1:47–56.
Tafdrup, S. 1995. Viable energy production and waste recycling from anaerobic digestion of manure and other biomass materials. Biomass and Bioenergy 9(1):303-314.
USDA NASS. 2017. 2017 Census of Agriculture. AC-17-A-51. https://www.nass.usda.gov/Publications/AgCensus/2017/Full_Report/Volume_1,_Chapter_1_US/usv1.pdf.
USDA NRCS (Natural Resources Conservation Service). 1992. Agricultural Waste Management Handbook.
USDA NRCS. 1995. Animal Manure Management. RCA Issue Brief No. 7. December. https://www.nrcs.usda.gov/wps/portal/nrcs/detail/null/?cid=nrcs143_014211.
USDA NRCS. 2011. National Agronomy Manual, Part 503, Crop Production. https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1043208.pdf.
USDA NRCS. 2012. General Manual, Title 190, Part 402, Nutrient Management. https://directives.sc.egov.usda.gov/viewerFS.aspx?hid=27119.
Wagena, M. B., and Z. M. Easton. 2018. Conservation practices can help mitigate the impact of climate change. Science of the Total Environment 635:132–143.
WAC (Watershed Agricultural Council). 2016. 2016 Strategic Plan. August.
WAC. 2017. Filtration Avoidance Determination Recommendations.
WAP (Watershed Agricultural Program). 2018. 2017 Precision Feed Management Program, Nutrient and Economic Impact Report. November.
Wetz, M., and D. Yaskowitz. 2013. An ‘extreme’ future for estuaries? Effects of extreme climatic events on estuarine water quality and ecology. Marine Pollution Bulletin 69(1–2):7-18.
Wiggs, G., Baird, A. and Atherton, R. 2004. The dynamic effects of moisture on the entrainment and transport of sand by wind. Geomorphology 59(1):13-30.
Woolf, D., J. E. Amonette, F. A. Street-Perrott, J. Lehmann, and S. Joseph. 2010. Sustainable biochar to mitigate global climate change. Nature Communications 1:56.
World Biogas Association. 2017. World Biogas Association’s Study Tour to the Netherlands, September 12. http://www.acceleratio.eu/world-biogas-study-tour-netherlands/.
This page intentionally left blank.