w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/watres

Observer rating of recreational use in wadeable streams of New York State, USA: Implications for nutrient criteria development Alexander J. Smith*, Brian T. Duffy, Margaret A. Novak New York State Department of Environmental Conservation, 425 Jordan Road, Troy, NY 12180, USA

article info

abstract

Article history:

Like most other States and Tribes in the United States, New York State has been working

Received 13 June 2014

with the United States Environmental Protection Agency to develop numeric nutrient

Received in revised form

criteria. These criteria are to protect water use such as drinking water supply, aquatic life,

13 November 2014

and recreation. Although extensive research exists related to the effects of eutrophication

Accepted 14 November 2014

on human health and aquatic life, limited information is available on perceived impair-

Available online 24 November 2014

ment of recreational opportunities in rivers and streams. We present an approach to assess impacts to recreation using information collected by New York State's (NYS) monitoring

Keywords:

program. This approach involved a questionnaire adapted from lake management surveys

User perception

in which field crews rated their perceptions of recreational ability at each site. The ratings

Observer rating

were then used to assess the relationship between perceived impact to recreational use

Recreational use

and water quality. We include in our analyses the primary nutrient criteria variables total

Nutrient criteria

phosphorus (TP), total nitrogen (TN), suspended chlorophyll-a (SChl-a), and turbidity (Tb), as well as biological condition (benthic macroinvertebrate community assessment). We sampled 203 wadeable stream locations throughout NYS between July and September 2008 e2012. Field crews ranked most locations as having “Minor aesthetic problems,” but still considered them excellent for both primary (34%) and secondary (37%) contact recreation. Field crew rankings of recreational ability coincided with a gradient of nutrients (TP and TN), SChl-a, and Tb concentration. Logistic regression models were developed that identified significant predictors affecting field crew decisions about recreation. These included water clarity, periphyton cover, and odor. Analysis of variance using NYS's multimetric assessment of biological condition and a nutrient specific community metric suggest significant differences in metric scores among recreational use categories. These results indicate correlation of impairment of recreational use with impairment of aquatic life use from nutrient enrichment. The results of this investigation will be used to help establish nutrient endpoints for the protection of recreation in NYS streams and rivers. Published by Elsevier Ltd.

* Corresponding author. Tel.: þ1 518 285 5627. E-mail addresses: [email protected] (A.J. Smith), [email protected] (B.T. Duffy), [email protected] (M.A. Novak). http://dx.doi.org/10.1016/j.watres.2014.11.022 0043-1354/Published by Elsevier Ltd.

196

1.

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

Introduction

The United States' Federal Clean Water Act (1972) provides the authority for states to develop water quality standards protective of water body designated uses as defined in Section 303(c) (2) (A). These uses include public water supply, propagation of fish, shellfish, and wildlife, recreation, and agriculture (Congress, 1972). Water resource scientists throughout the United States are working to develop numeric water quality standards for nitrogen and phosphorus protective of these uses. This follows recommendations by the United States Environmental Protection Agency (USEPA) in their national strategy for the development of nutrient criteria (USEPA, 1998). The USEPA's call for nutrient criteria stems from increased impairment of surface waters in the United States from eutrophication, the primary source of which is nitrogen and phosphorus from both point and nonpoint sources (USEPA, 2000). In New York State (NYS) alone, nutrients are the source of more than half of all stream and river water quality impact (Bode et al., 2004). Extensive research has been conducted relating the impacts of excessive nutrient concentrations on aquatic life (Dodds et al., 1998, 1997; Dodds and Welch, 2000; Havens, 2003; Sheeder and Evans, 2004; Smith et al., 2007; Stevenson et al., 2006; Wang et al., 2007) and to a lesser extent human health (Arruda and Fromm, 1989; Codd, 2000; Downing et al., 2001; Palmstrom et al., 1988), and the impairment of recreational uses and aesthetics in lakes (Ditton and Goodale, 1973; Heiskary and Walker, 1988; Smeltzer and Heiskary, 1990; Wilson and Walker, 1989). However, limited information pertaining to impairment of recreational opportunities specific to rivers and streams exists, and has implications for developing nutrient criteria (Suplee et al., 2009). As numeric water quality standards for nutrients are developed it will be important to ensure that all Clean Water Act uses are protected including recreation in rivers and streams. To do so, water quality managers will need to focus resources on connecting impacts to recreation with nutrient over-enrichment in ways that parallel the attention paid to aquatic life, human health, and recreational use in lakes. The perceived quality of both the scenic and recreational environment has typically been studied through the use of visual research and observer rating surveys involving some portion of the general public. These methods have been used extensively in various applications from estimating the importance of scenic beauty to assessing the implications of forest practices and the value of campground cleanliness (Brush, 1976; Manning and Freimund, 2004). The use of public perception data in the management of water quality and water resource policy has typically been associated with lake and reservoir management. In this application public perception data are often collected through the deployment of surveys to directly measure the opinion of the user community (Butler and Redfield, 1991). Surveys range from a simple one or two question form about water body condition or recreational ability (Smeltzer and Heiskary, 1990) to more complex questionnaires requiring the ranking of individual forms of pollution, some using visual research methods which include evaluation of site photographs (House, 1996; Manning

and Freimund, 2004; Smith et al., 1991; Suplee et al., 2009). Water quality managers then use the survey responses in an attempt to relate traditional measures of water quality, such as clarity, chlorophyll-a, or phosphorus to assess impacts to recreation (Smith et al., 1991). The use of public perception survey to gage perceived recreational quality is more complex than for lakes. In New York, streams have boundaries which stretch on for miles, are more varied, and lack the dense populations of the lakeshore community or organized lake association. This can make identification of the user population difficult and hinder large scale statewide deployment of user opinion surveys. However, given the current census coverage available in the United States, these hindrances are no longer insurmountable. For example, Suplee et al. (2009) successfully deployed a user perception survey using a Centralized Voter File in Montana. Similar to many investigations reliant on survey data, the effect of nonrespondents on the final outcome tends to be the greatest source of error (Butler and Redfield, 1991; Suplee et al., 2009). However, in some cases user opinion surveys of water quality include in-person interviews or on-site survey visits (House, 1996; Smith et al., 1991; Suplee et al., 2009), which may help reduce error introduced through non-response. Given the difficulty and resources involved in performing a statewide public perception survey of stream and river water quality, we suggest an alternate approach based on data from field crews in this monitoring program. This approach involves the rating of recreational ability at wadeable stream sites by water quality monitoring program field crews. The rating procedure uses a routine questionnaire similar to other public perception polls and adapted from public perception surveys used in lake management (Heiskary and Walker, 1988; Smeltzer and Heiskary, 1990) with only slight modifications. We suggest that recreational use ratings generated by field crews are robust and can be directly associated with changes in water quality conditions. However, we acknowledge that this method of data collection is not without some weakness. The same can also be said for obtaining recreational use data from public opinion surveys, none of which is insurmountable. For example, public opinion surveys are susceptible to 1) responses from individuals regionally acclimated to specific water quality conditions (Heiskary and Walker, 1988) 2) difficult randomization of sample design within specific resource user groups (Smeltzer and Heiskary, 1990) 3) effects of demographics on responses (House, 1996) and 4) high nonresponse rates (Butler and Redfield, 1991; Suplee et al., 2009). Most of these drawbacks can be addressed through simple characterization of the target population being surveyed so that results are framed within the proper context and not applied outside these bounds though some issues can be addressed directly. For example, high non-response rates can be addressed through follow-up in-person interviews (Butler and Redfield, 1991; Suplee et al., 2009). We think the use of ratings generated by field crews is less likely to be susceptible some of these opinion survey problems. Field crews are familiar with the diversity of water quality conditions statewide, meaning their ratings of recreational use will provide a good foundation for statewide comparison of use attainment. However, this is also a drawback since field crew's detailed knowledge of water quality

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

conditions could result in significant differences of opinion compared to the actual public. Randomization of survey design using field crews is controlled through the design of the monitoring program. Many states such as New York now employ a random probabilistic survey design for at least some component of their monitoring program. Similar to a the public opinion survey, demographics may play a role in the results from field crews if there is a range of ages and both sexes are represented. Non-response is not usually an issue for field crew generated ratings since the collection of the information is integrated into routine site assessments. In the absence of public opinion survey data ratings provide useful information for beginning to develop water quality standards in streams and rivers protective of recreational use. In fact much of the information gathered in this investigation suggest that managing other water quality endpoints (such as existing standards or biology) may result in adequate protection of recreational uses. Because the rating questionnaire is integrated into a routine monitoring program, the information is coupled with detailed water quality and biological monitoring data statewide. The association of nutrient and chlorophyll-a data with this information can be used in developing numeric nutrient criteria. To that end we evaluated data on field crew ratings of recreational use and its relationship with total phosphorus, total nitrogen, suspended chlorophyll-a, turbidity, and the response of the benthic macroinvertebrate community. We hypothesized that field crews would rank recreational use impairment in a manner concordant with a specific range of nutrient variables similar to findings from studies of streams in Montana (Suplee et al., 2009) or lakes in Vermont and Minnesota (Heiskary and Walker, 1988; Smeltzer and Heiskary, 1990) where chlorophyll-a, phosphorus, secchi depth, and nuisance algal blooms were related to public perception surveys. In addition,

197

just as House (1996) reported observations of oil, trash, or sewage having the greatest impact on survey respondent's answers, we predicted that field crews would identify several specific observational elements that routinely influenced their ratings of recreational ability. Using the results we propose numeric values of total phosphorus, total nitrogen, and chlorophyll-a as nutrient criteria for the protection of recreational uses in NYS.

2.

Material and methods

To assess the relationship between ratings of recreational use and water quality, including nutrients and biological condition, we sampled 203 wadeable streams throughout NYS between July and September 2008e2012 (Fig. 1). Approximately half of the dataset (100 sites) was reported on previously in the development of regional nutrient thresholds in wadeable streams of NYS for the protection of aquatic life (Smith et al., 2013). A complete discussion of the water quality variables collected e water chemistry (primarily nutrients), biological (macroinvertebrate) communities, and physical habitat measurements is contained in Smith et al. (2013). Relevant elements of this dataset are discussed here. The remaining 103 sites were sampled as part of the NYS Department of Environmental Conservation's (NYSDEC) ambient water quality monitoring program in which water chemistry, biological community, and field crew ratings of recreational use were also collected. Sampling locations were distributed across NYS to represent a gradient of nutrient conditions (Fig. 1) in wadeable streams. Selection of sampling locations used percent land cover data, historical water chemical data, and biological community assessment information. Approximately 10% of

Fig. 1 e Map of the 203 different sampling locations where observer ratings were completed by field crews across New York State. Sampling occurred between July and September 2008e2012.

198

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

the entire population of sites was considered reference, defined by having 75% natural cover in the upstream watershed, background nutrient levels based on previous water sampling, and historical biological condition assessments indicating naturally occurring, non-impacted communities. This ensured that a portion of the dataset represented the least-disturbed and/or best-attainable condition (Reynoldson et al., 1997). Remaining sites had a range of natural cover in their upstream watershed and represented a gradient of nutrient and biological conditions outside of the reference condition boundaries. As noted in Smith et al. (2013) similar methods were employed during previous nutrient criteria studies and resulted in adequate representation of streams and rivers with different nutrient and community status (Smith and Tran, 2010). The selection process resulted in a population of wadeable streams of similar size and type along a gradient of disturbance. Field data showed stream types were similar all of which were hard bottom high gradient riffle habitat streams with interquartile ranges for drainage area of 31e181 km2, width of 5e15 m, depth of 0.1e0.2 m, current speed of 50e80 cm/s, percent rock substrate of 10e25%, rubble substrate 20e35%, gravel 20e35%, sand 10e20%, and silt 7e15%. Geographic information system data showed forest cover of 46e80 % dominated by eastern hardwood, specific conductance of 146e560 msiemen/cm, temperature of 18e22  C, and dissolved oxygen of 8e10 mg/L. At each sampling location a questionnaire for rating recreational ability was completed by a field crew member of the NYSDEC water quality monitoring program. The questionnaire was used to rate their perspective on whether or not the water body supported the recreational use it was meant to sustain (Fig. 2). Additional details and copies of the questionnaire can be found in the Standard Operating Procedure: Biological Monitoring of Surface Waters in New York State (NYSDEC, 2012). The survey assessed field crew perceptions of impacts to primary and secondary contact recreation. New York State's Environmental Conservation Law, Part x700.1 defines primary contact as “recreational activities where the human body may come in direct contact with raw water to the point of complete body submergence. Primary contact recreation includes swimming, diving, water skiing, skin diving and surfing,” and secondary contact is defined as “recreational activities where contact with the water is minimal and where ingestion of the water is not probable. Secondary contact recreation includes “fishing and boating” (NYS, 1967). The questionnaire was conducted by one field crew member at each sampling location. There were 16 different field crew members, comprising 19 unique field crews evaluating recreational use impacts during the course of the fiveyear data collection period. The distribution of overall water quality conditions (biology and nutrients) among unique field crews was assessed to determine the similarity of sampling site populations visited by each field crew member. Overall water quality condition was consistently distributed among field crews across both datasets used in this investigation, with all field crews assigned a higher proportion of good water quality locations. This was a function of the datasets available and the distribution of sites along a gradient of water quality condition. The questionnaire was completed upon arrival at each sampling location prior to additional data collection,

thereby limiting field crew bias from knowledge of other water quality parameters. The primary component of the questionnaire was a pair of questions meant to facilitate rating impacts to both primary and secondary contact recreation. The questions were multiple choice and offered a set of answers ranging from “Beautiful, could not be nicer” to “Awful, recreation is impossible.” There were six possible answers for each question, each of which was renumbered sequentially from 1 to 6 for use in statistical analyses. The field crew member filling out the form circled one answer for each question (Fig. 2). The last portion of the survey dealt with ranking the presence of specific observational variables affecting his or her decisions in the first two questions. The variables were each listed on the questionnaire with a scale from 0 (natural) e 10 (highly disturbed). The field crew member circled one value on the scale for each variable. The variables evaluated were water clarity, phytoplankton, periphyton cover, macrophyte cover, odor, trash and discharges/pipes (Fig. 2). The primary survey questions and numeric ranking of each observational variable were used in evaluating the relationship between water chemistries, biological integrity, and ratings of recreational use ability. The method of collecting water chemical samples varied slightly between the two different datasets used in the investigation. However, in both datasets, water chemical samples were collected during base flow conditions and were accompanied by in-situ measurements of temperature, pH, specific conductance, and dissolved oxygen using a calibrated field meter (YSI 556 multi-probe meter; Yellow Springs Instruments, Yellow Springs, Ohio). Water chemical samples were collected using a DH-81 depth integrated sampler and 10% of which were accompanied by field and laboratory blanks for quality control. In the dataset provided by Smith et al. (2013), water chemical samples were collected once on the same date and time the rating questionnaire was completed and the biological sample was collected. Water chemical samples from the NYSDEC's ambient monitoring program were collected up to six times between the months of June and September in the same year the questionnaire was completed and the biological samples were collected. Because multiple water chemical samples existed for each site we used the average of these events for further analysis and reporting. For the purposes of this investigation we focus on the results of the primary nutrients total phosphorus (TP), and total nitrogen (TN), as well as suspended chlorophyll-a (SChl-a), and turbidity (Tb). Detection limits for both datasets were comparable because the similar analytical methods were used (TP: 2 mg/L, TN: 100 mg/L, SChl-a: 0.02 mg/L, Tb: 0.3 NTU). SChl-a was not collected as part of the NYSDEC's ambient monitoring program therefore results of the analyses are based on n ¼ 100, rather than n ¼ 203 as they are for TP, TN, and Tb. Water chemical samples were shipped on ice to a contract laboratory for analysis. A full suite of water chemical variables was analyzed as part of the larger projects these datasets supported. For details on the full suite collected the reader is referred to the article “Regional nutrient thresholds in wadeable streams of New York State protective of aquatic life” (Smith et al., 2013) and the “Rotating Integrated Basins Study (RIBS) Water Sampling Standard Operating Procedures” (NYSDEC, April 2010). Reasons for focusing on these chemical

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

199

Fig. 2 e The use perception questionnaire provided to each field crew conducting routine water quality sampling during this investigation. The first two multiple choice questions address perceptions of primary and secondary contact recreation ability respectively. The remainder of the questionnaire focuses on identifying which variables may have most influenced the field crew member's decision in the first two questions. The form was based on previous surveys used in lake and reservoir management in Vermont, Minnesota, and New York.

200

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

variables include their prominence in the USEPA national strategy for nutrient criteria development and their direct relationship with algae (benthic and suspended) and macrophyte growth (Raschke, 1993; Smith et al., 1999; Van Nieuwenhuyse and Jones, 1996), as well as previous accounts of impacts to recreation use (Burden and Malone, 1987; Heiskary and Walker, 1988; Suplee et al., 2009). We collected benthic macroinvertebrate samples once at each sampling location between July and September. Methods followed those outlined in the Standard Operating Procedure: Biological Monitoring of Surface Waters in New York State (NYSDEC, 2012). Samples were collected using a traveling kick technique to disrupt the stream bottom for a five-minute period over a five-meter diagonal transect in the dominant riffle at each sampling location. The net used was a 22.86 cm  45.72 cm aquatic dip net with a 0.8 mm  0.9 mm mesh opening. In the field, all sample contents were sieved using a U.S. no. 30 standard sieve, transferred to a quart jar and then preserved with 95% ethyl alcohol. Samples were sent to a contract laboratory specializing in the taxonomy of freshwater benthic macroinvertebrates, or processed inhouse by the NYSDEC Stream Biomonitoring Unit (SBU). Benthic macroinvertebrate sample processing included the enumeration of 100eorganism subsamples and identification of specimens to the lowest possible taxonomic resolution (i.e. genus or species). Benthic macroinvertebrate community data were synthesized into NYSDEC's Biological Assessment Profile score (BAP). The BAP is a multimetric used to distill macroinvertebrate data into a single score for routine use in the biological assessment of water quality, specifically for the purposes of fulfilling NYSDEC's reporting requirements to USEPA. The BAP places macroinvertebrate communities on a scale of 0e10 with a score of 7.5e10 indicating non-impacted conditions, 5e7.5 slightly impacted, 2.5e5.0 moderately impacted, and 0e2.5 severely impacted (NYSDEC, 2012; Smith et al., 2007, 2013). The BAP score is the mean of five individual component metrics which have been converted to a common scale (NYSDEC, 2012; Smith et al., 2007, 2013). These individual metrics include: Species richness (Spp.), Ephemeroptera/Plecoptera/Trichoptera richness (EPT) (Lenat, 1988), Hilsenhoff's Biotic Index (HBI) (Hilsenhoff, 1987), Percent Model Affinity (PMA) (Novak and Bode, 1992), and Nutrient Biotic Index (NBI) (Smith et al., 2007). The relationship between field crew ratings of recreational ability and water chemical variables was explored using nonparametric analysis of variance and graphical representation in box and whisker plots. This comparison helped provide understanding as to whether or not managing for specific levels of nutrient and biological condition would result in the protection of recreational use. KruskaleWallis rank sum tests were performed using the primary nutrients (TP and TN), SChl-a, and Tb with field crew responses to ranking primary and secondary contact recreation as categories for comparison. A multiple comparisons test was used to identify specific categorical differences in water chemical values. We used the functions Kruskal test and Kruskalmc (a ¼ 0.05) in R (version 2.12) (R-development Team, 2010). Using this information we identified the recreational use category representing the most significant change in water chemical concentrations for both

recreation types. The median TP, TN, SChl-a, and Tb concentrations for each recreational use category was calculated. Median values were determined most suitable as the measure of central tendency for each of the categories because of skewed distribution of the data. We present these values as potential thresholds for nutrient criteria for the protection of recreational use. The observational variables affecting field crew member's decision in the rating of recreational use ability (Fig. 2) were examined for correlation with water chemical variables. This was conducted using Spearman Rank Correlation, rcorr function in R 2.12 (R-development Team, 2010). A matrix of scatter plots was generated to identify variables associated with water chemistries but with curvilinear relationships. Variables affecting recreational use ratings were reported if their correlation with TP, TN, SChl-a, and Tb concentrations exceeded r  0.4, a  0.05. To further elucidate which variables were most influential in determining impacts to recreational use we included the full set of observational variables from the questionnaire in a logistic regression model. These variables are water clarity, phytoplankton, periphyton, macroalgae, odor, trash, and number of pipes/discharges. In the KruskaleWallis rank sum tests using water chemical variables we found the recreational use rating category of “Slightly impacted” to represent the transition point between significant differences in water chemical concentrations. Based on this information we recoded the recreational use ranking variable for primary and secondary contact recreation to reflect a binary condition of “Not Impaired” (questionnaire ranking of “Beautiful, could not be nicer”, “Minor aesthetic problems,” and “Slightly impacted”) and “Impaired” (“Desire to participate substantially reduced” and “Awful, contact recreation impossible”). Using the general linear model function (glm), type “binomial” in R (version 2.12) (R-development Team, 2010) the results of this logistic regression provided a model with multiple significant predictor variables with associated odds ratios. The outcomes for the coefficients in the regression models suggest the best predictors related to the transition to a state of impaired recreational use (Kutner et al., 2004). In addition to comparison between recreational use ratings and water chemical variables we also explored the relationship between recreational use and biological community response. Using macroinvertebrate BAP scores, we conducted a similar KruskaleWallis rank sum test to that performed between perception data and water chemical variables. With the emphasis on nutrients and the development of nutrient criteria we applied this rank comparison test using the NBI as well as the BAP because of its demonstrated response to eutrophication (Smith et al., 2007, 2013; Smith and Tran, 2010). Average scores for both the BAP and NBI from the different recreational use categories are provided.

3.

Results

Based on field crew ratings of recreational use the category “Minor aesthetic problems,” contained the highest percentages of sites for both primary (34%) and secondary (37%) contact recreation. The remaining breakdown of the

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

recreational use rating categories was as follows; for primary contact recreation 11% “Beautiful, could not be nicer”, 26% “Slightly impacted”, 22% “Desire to participate substantially reduced”, 7% “Awful, contact recreation impossible.” For secondary contact recreation a greater percentage of locations were categorized as “Beautiful, could not be nicer” (23%) and fewer were categorized as “Slightly impacted” (19%). Field crews consistently ranked streams based on their perceived ability to recreate in a manner coinciding with a gradient of nutrients (TP and TN), SChl-a, and Tb concentration (Fig. 3). The highest concentration of each water chemical variable was associated with the worst ranked recreational use categories including “Desire to participate substantially reduced” and “Awful, contact recreation impossible.” Water chemical concentrations were lower and less variable within the recreational use rankings of “Beautiful, could not be nicer,” “Minor aesthetic problems,” and “Slightly impacted” (Fig. 3). The results of the KruskaleWallis rank sum tests further indicate field crews were able to categorize sites along a gradient of nutrient concentration and recreational use. Significant differences were identified between the recreational use rankings using the multiple comparisons test. In each of the comparisons evaluated (TP, TN, SChl-a, and Tb) the categories “Minor aesthetic problems” and “Desire to participate substantially reduced” were consistently significantly different from each other (TP, TN, Tb, n ¼ 203, df ¼ 4, p < 0.05; SChl-a, n ¼ 100, df ¼ 4, p < 0.05). This held true regardless of whether or not the comparison was made using rankings of primary or secondary contact recreation. These results suggest that the greatest change in chemical variable concentration occurred between these two categories. The category that falls between these two is recreational use “Slightly impacted.” To prevent recreational use from entering into a state of recreational ability “substantially reduced”, we propose the median concentration of TP, TN, SChl-a, and Tb from the “Slightly impacted” category as thresholds not to be exceeded. These chemical variable thresholds should prevent in-stream conditions of algal growth or reduced clarity that causes users to rank sites as “Desire to participate substantially reduced” or “Awful, contact recreation impossible.” Since these categories tend to represent the highest chemical variable concentration for the remainder of this investigation we consider sites ranked as such to be impaired for recreational use. Median concentrations of water chemical variables are provided for each of the recreational use categories (Table 1). The median values for the category of “Slightly impacted” indicate only slightly higher concentrations for secondary contact recreation from this same category for primary contact recreation (Table 1). Field crews also ranked the observational variables describing the condition of each location in a way that correlated with the results of measured water chemical variables. Of the six different observational variables evaluated (Fig. 2) Spearman rank correlation suggested the following had the most significant relationships with water chemistries: water clarity was correlated with TP (r ¼ 0.46, p < 0.0001), Tb (r ¼ 0.62, p < 0.0001), and SChl-a (r ¼ 0.59, p < 0.0001), phytoplankton was correlated with SChl-a only (r ¼ 0.50, p < 0.0001), periphyton cover was correlated with TP (r ¼ 0.46, p < 0.0001), TN (r ¼ 0.42, p < 0.0001), Tb (r ¼ 0.40, p < 0.0001), and SChl-a

201

(r ¼ 0.49, p < 0.0001), and odor was correlated with TP (r ¼ 0.40, p < 0.0001), TN (r ¼ 0.44, p < 0.0001), and SChl-a (r ¼ 0.45, p < 0.0001). The logistic regression models we developed identified several significant predictors of impaired recreational use (i.e. rating of sites as “Desire to participate substantially reduced” or “Awful, contact recreation impossible”). For both primary and secondary contact recreation, water clarity, periphyton growth, and odor were significant predictors of impaired recreational use (Table 2, Fig. 4). In addition, the presence of pipes and trash were also significant predictors, but weaker, and were restricted to secondary contact recreation. Regardless, odor, trash, and pipes were not nearly as strong predictors of the transition to impaired recreational use compared to both water clarity and periphyton (Table 2). Results of overall model fit were significant for both primary and secondary contact recreation with chi-square of 124.5, 7 df, p < 0.0001, and chi-square of 95.8, 7 df, p < 0.0001, respectively (Table 2). The odds ratios associated with the significant predictors of impaired recreational use suggest the probability of rating recreation as “impaired” due to one unit increase in the predictor variable. For example, for a one unit decrease in water clarity, the odds of a field crew member classifying a site as “impaired” for primary contact recreation increases by a factor of 1.6 when all other predictors are held constant (Table 2). The odds ratios corresponding to each of the significant model predictors is provided in Table 2. The interpretation just given holds for each. Similar to the concordance of ratings of recreational ability with gradients of water chemical variables, field crews also rated capacity to support recreation in a way that corresponded with a gradient of biological condition. The results of the KruskaleWallis rank sum test performed using both NYS's multimetric assessment of water quality (BAP) and nutrient biotic index (NBI) suggest significant differences in the scores of these metrics between recreational use categories (Fig. 5). Differences were again most defined between the categories of “Minor aesthetic problems” and “Desire to participate substantially reduced” (Primary Contact: BAP and NBI, n ¼ 203, df ¼ 4, p < 0.05; Secondary Contact: BAP and NBI, n ¼ 203, df ¼ 4, p < 0.05). The median scores for both the BAP and NBI between the primary contact recreation categories of “Beautiful, could not be nicer” and “Minor aesthetic problems” ranged from 8.1 to 7.8 and 5.3e5.4, and from 7.9 to 7.5 and 5.3e5.7 for secondary contact recreation. For the categories of “Desire to participate substantially reduced” and “Awful, contact recreation impossible” the median scores ranged between 5.5 and 4.3 for the BAP and 7.0 consistently for the NBI based on primary contact categories and they ranged from 5.3 to 4.1 for the BAP and 7.0e7.2 for the NBI based on secondary contact recreation categories (Table 1). These results represent a decline of overall water quality measured by the BAP score and an increase in nutrient enrichment measured by the NBI. The change in BAP scores represented a shift from nonor slightly impacted biological condition at sites rated “Beautiful, could not be nicer” and “Minor aesthetic problems” to impaired biological conditions (BAP  5.0) for the categories “Desire to participate substantially reduced” and “Awful, contact recreation impossible”. These results also suggest a similar worsening in the degree of eutrophication as

202

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

203

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

Table 1 e Median water chemical values and benthic macroinvertebrate community metrics of samples categorized by field crew responses to a use perception survey. Chemical variables presented are total phosphorus (TP), total nitrogen (TN), suspended chlorophyll-a (SChl-a), and turbidity (Tb). Benthic macroinvertebrate community metrics include NYS's multimetric, the Biological Assessment Profile (BAP) score and the eutrophication specific Nutrient Biotic Index (NBI). Perception Category

Primary contact recreation

Secondary contact recreation

TP ug/L TN ug/L SChl-a ug/L Tb NTU BAP NBI TP ug/L TN ug/L SChl-a ug/L Tb NTU BAP NBI Beautiful Minor Slightly Impacted Substantially Reduced Awful

18 16 26 36 50

539 440 709 965 1756

1.2 1.3 1.8 5.2 4.2

2.7 2.3 4.0 5.0 8.1

8.1 7.8 6.5 5.5 4.3

5.3 5.4 6.4 7.0 7.0

16 18 29 50 59

519 532 714 1042 1756

1.1 1.9 2.1 3.4 16.0

2.9 2.4 4.3 4.7 8.5

7.9 7.5 5.9 5.3 4.1

5.3 5.7 6.6 7.2 7.0

Table 2 e Results of logistic regression models. Field crew responses to survey questions about ability to participate in primary and secondary contact recreation were transformed to a binomial response of “not impaired” or “impaired” based on results of analysis of variance. Individual predictors of the binomial response were identified that increase the odds of resulting in an impaired assessment of recreational ability. Coefficient

Estimate

Primary contact recreation Clarity Phytoplankton Periphyton Macroalgae Odor Trash Pipes Null deviance Residual deviance Chi-square Secondary contact recreation Clarity Phytoplankton Periphyton Macroalgae Odor Trash Pipes Null deviance Residual deviance Chi-square

SE

Z-value

P-value

0.45971 0.0729 0.39576 0.04467 0.63390 0.53173 0.1075 245.8 121.3 124.5

0.11901 0.37924 0.09859 0.15139 0.24270 0.17829 0.35989 df df df

3.863 0.192 4.014 0.295 2.612 2.982 0.299 201 194 7

0.0001 0.8476 0.0001 0.7679 0.0090 0.0029 0.7652

1.6

60

1.5

50

1.9 1.7

90 70

p-value

65 mg/L (Smith et al., 2007). Therefore, for both primary and secondary contact recreation the results suggest not only impaired recreation but also impaired biological condition that is most likely the result of eutrophication with measured TP over 26 mg/L and TN over 700 mg/L (Table 2).

5.

Conclusions

Observer ratings of recreational use are a useful and meaningful component of assessing the condition of waters for recreational and aesthetic uses. Unfortunately in rivers and streams we are only just now recognizing the utility of this type of information. In the realm of developing nutrient criteria, user perception data (expert or non-expert derived) are likely to have some of its greatest impact on water resource policy. The conditions our investigation and others have shown to lead to undesirable ratings of recreational use are common in highly eutrophic systems. For example, poor water clarity from sediment and suspended algae, or abundant algal growth on bottom substrates. Correlation of water chemical variables and biological condition with ratings of recreational use also provides useful benchmarks for developing other water quality criteria. The ability to relate water chemical variables to observer ratings means we have a method for targeting specific chemical concentrations for the purpose of protecting recreation. The method of collecting observer ratings data on recreational use discussed here provides a baseline for statewide comparison. However, we acknowledge that these conclusions are drawn from field crew responses to a questionnaire, and therefore caution is needed in assuming these responses are at all representative of the general public. In the future it will be beneficial, now that the utility of this information has been shown, to consider a statewide public opinion poll for stream and river recreation use. NYS could then integrate user perception surveys into routine assessments as part of their 305(b) and 303(d) reporting. In addition, this type of survey could be integrated as part of the National Aquatic Resource Surveys conducted by the USEPA. User perception surveys would be a substantial accompaniment to these random probabilistic surveys of general water quality condition nationally.

Acknowledgments The authors would like to thank the incredible staff of the NYSDEC Stream Biomonitoring Unit for their effort in this project including Diana Heitzman and Jeff Lojpersberger; Roger Thomas, David Velinsky, and Sylvan Klein of the Philadelphia Academy of Natural Sciences for collecting and recording field measurements; J. Kelly Nolan and the staff of Watershed Assessment Associates for their expertise in processing biological samples; and the United States Environmental Protection Agency for providing funding for this project among other nutrient criteria development work over the past 10 years.

208

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

references

Arcury, T.A., 1990. Environmental attitude and environmental knowledge. Hum. Organ. 49, 300e304. Arcury, T.A., Johnson, T.P., 1987. Public environmental knowledge: a statewide survey. J. Environ. Educ. 18, 31e37. Arruda, J.A., Fromm, C.H., 1989. Relationships among trihalomethane formation potential, organic carbon and lake enrichment. Environ. Pollut. 61, 199e209. Bode, R.W., Novak, M.A., Abele, L.E., Heitzman, D.L., Smith, A.J., 2004. 30 Year Trends in Water Quality of Rivers and Streams in New York State Based on Macroinvertebrate Data. New York State Department of Environmental Conservation, Division of Water, Albany, NY. Bord, R.J., Fisher, A., O'Connor, R.E., 1998. Public perceptions of global warming: United States and international perspectives. Clim. Res. 11, 75e84. Brush, R.O., 1976. Perceived quality of scenic and recreational environments. In: Craik, K.H., et al. (Eds.), Perceiving Environmental Quality. Plenum Press, New York, pp. 47e58. Burden, D.G., Malone, R.F., 1987. A classification of freshwater Louisiana lakes based on water quality and user perception data. Environ. Monit. Assess. 9, 179e193. Butler, L.T., Redfield, G.W., 1991. The restion (Virginia) Lake use survey: public perceptions and management implications. J. Am. Water Resour. Assoc. 27, 603e610. Codd, G.A., 2000. Cyanobacterial toxins, the perception of water quality, and the prioritisation of eutrophication control. Ecol. Eng. 16, 51e60. Congress, U.S., 1972. Federal Clean Water Act (Washington, D.C.). Ditton, R.B., Goodale, T.L., 1973. Water quality perception and the recreational uses of Green Bay, Lake Michigan. Water Resour. Res. 9, 569e579. Dodds, W.K., Jones, J.R., Welch, E.B., 1998. Suggested classification of stream trophic state: distributions of temperate stream types by chlorophyll, total nitrogen, and phosphorus. Water Res. 32, 1455e1462. Dodds, W.K., Smith, V.H., Zander, B., 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: a case study of the Clark Fork River. Water Res. 31, 1738e1750. Dodds, W.K., Welch, E.B., 2000. Establishing nutrient criteria in streams. J. North Am. Benthol. Soc. 19, 186e196. Downing, J.A., Watson, S.B., McCauley, E., 2001. Predicting cyanobacteria dominance in lakes. Can. J. Fish. Aquat. Sci. 58, 1905e1908. Dunlap, R.E., 1991. Trends in public opinion toward environmental issues: 1965e1990. Soc. Nat. Resour. 4, 285e312. Havens, K.E., 2003. Phosphorus-algal bloom relationships in large lakes of South Florida: Implications for establishing nutrient criteria. Lake Reserv. Manag. 19, 222e228. Heiskary, S.A., Walker, W.W., 1988. Developing phosphorus criteria for Minnesota lakes. Lake Reserv. Manag. 4, 1e9. Hilsenhoff, W.L., 1987. An improved biotic index of organic stream pollution. Gt. Lakes. Entomol. 20, 31e40. House, M., Fordham, M., 1997. Public perceptions of river corridors and attitudes towards river works. Landsc. Res. 22, 25e44. House, M.A., 1996. Public perception and water quality management. Water Sci. Technol. 34, 25e32. House, M.A., Sangster, E., 1991. Public perception of RiverCorridor management. Water Environ. J. 5, 312e316. € ki, S., Savolainen, R., 1984. The scenic value of the forest Kelloma landscape as assessed in the field and the laboratory. Landsc. Plan. 11, 97e107.

Kutner, M.H., Nachtsheim, C., Neter, J., 2004. Applied Linear Regression Models, Fourth ed. ed. McGraw-Hill Irwin, New York, NY. Lenat, D.R., 1988. Water quality assessment of streams using a qualitative collection method for benthic macroinvertebrates. J. North Am. Benthol. Soc. 7, 222e233. Lowe, G.D., Pinhey, T.K., Grimes, M.D., 1980. Public support for environmental protection: new evidence from national surveys. Pac. Sociol. Rev. 423e445. Manning, R., Freimund, W., 2004. Use of visual research methods to measure standards of quality for parks and outdoor recreation. J. Leis. Res. 36, 557e579. Novak, M.A., Bode, R.W., 1992. Percent model affinity: a new measure of macroinvertebrate community composition. J. North Am. Benthol. Soc. 11, 80e85. NYS, 1967. Environmental Conservation Law. Chapter X. New York State Department of Environmental Conservation, Division of Water. New York State Legislature, Albany, NY. NYSDEC, 2012. Standard Operating Procedure: Biological Monitoring of Surface Waters in New York State. New York State Department of Environmental Conservation, Division of Water, Albany, NY. NYSDEC, April 2010. Rotating Integrated Basins Study (RIBS) Water Sampling Standard Operating Procedures, Revision 1.0. New York State Department of Environmental Conservation, Division of Water, Albany, NY. Palmstrom, N., 2005. Development of regional nutrient criteria and implications for the regulated community. Proc. Water Environ. Fed. 2002, 532e540. Palmstrom, N.S., Carlson, R.E., Cooke, G.D., 1988. Potential links between eutrophication and the formation of carcinogens in drinking water. Lake Reserv. Manag. 4, 1e15. Raschke, R.L., 1993. Guideline for assessing and predicting eutrophication status of small Southeastern Piedmont impoundments. In: Hatcher, K.J. (Ed.), Georiga Water Resources Conference. Institute of Natural Resources, The University of Georgia, Athens, pp. 83e86. R-development Team, 2010. R: a Language and Environment for Statistical Computing, 2.12.0 Ed. R Foundation for Statistical Computing, Vienna, Austria. Reynoldson, T.B., Norris, R.H., Resh, V.H., Day, K.E., Rosenberg, D.M., 1997. The reference condition: a comparison of multimetric and multivariate approaches to assess waterquality impairment using benthic macroinvertebrates. J. North Am. Benthol. Soc. 16, 833e852. Rohm, C.M., Omernik, J.M., Woods, A.J., Stoddard, J.L., 2002. Regional characteristics of nutrient concentrations in streams and their application to nutrient criteria development. J. Am. Water Resour. Assoc. 38, 213e239. Rosenberg, D.M., Resh, V.H., 1993. Freshwater Biomonitoring and Benthic Macroinvertebrates. Chapman & Hall, New York, NY. Scott, D., Willits, F.K., 1994. Environmental attitudes and behavior a Pennsylvania survey. Environ. Behav. 26, 239e260. Sheeder, S.A., Evans, B.M., 2004. Estimating nutrient and sediment threshold criteria for biological impairment in Pennsylvania watersheds. J. Am. Water Resour. Assoc. 40, 881e888. Smeltzer, E., Heiskary, S.A., 1990. Analysis and applications of Lake user survey data. Lake Reserv. Manag. 6, 109e118. Smith, A.J., Bode, R.W., Kleppel, G.S., 2007. A nutrient biotic index (NBI) for use with benthic macroinvertebrate communities. Ecol. Indic. 7, 371e386. Smith, A.J., Thomas, R.L., Nolan, J.K., Velinsky, D.J., Klein, S., Duffy, B.T., 2013. Regional nutrient thresholds in wadeable streams of New York State protective of aquatic life. Ecol. Indic. 29, 455e467.

w a t e r r e s e a r c h 6 9 ( 2 0 1 5 ) 1 9 5 e2 0 9

Smith, A.J., Tran, C.P., 2010. A weight-of-evidence approach to define nutrient criteria protective of aquatic life in large rivers. J. North Am. Benthol. Soc. 29, 875e891. Smith, D.G., Cragg, A.M., Croker, G.F., 1991. Water clarity criteria for bathing waters based on user perception. J. Environ. Manag. 33, 285e299. Smith, V.H., Tilman, G.D., Nekola, J.C., 1999. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environ. Pollut. 100, 179e196. Stevenson, R.J., Rier, S.T., Riseng, C.M., Schultz, R.E., Wiley, M.J., 2006. Comparing effects of nutrients on algal biomass in streams in two regions with different disturbance regimes and with applications for developing nutrient criteria. Hydrobiologia 561, 149e165. Suplee, M.W., Varghese, A., Cleland, J., 2007. Developing nutrient criteria for streams: an evaluation of the frequency distribution method. J. Am. Water Resour. Assoc. 43, 453e471. Suplee, M.W., Watson, V., Teply, M., McKee, H., 2009. How green is too green? public opinion of what constitutes undesirable algae levels in Streams1. J. Am. Water Resour. Assoc. 45, 123e140. USEPA, 1998. National Strategy for the Development of Regional Nutrient Criteria. United States Environmental

209

Protection Agency, Office of Water, Washington, D.C. EPA 822-R-98e022. USEPA, 2000. Nutrient Criteria Technical Guidance Manual, Rivers and Streams. United States Environmental Protection Agency, Office of Water, Office of Science and Technology, Washington, D.C. EPA 822-B-00e002. Van Nieuwenhuyse, E.E., Jones, J.R., 1996. Phosphorus chlorophyll relationship in temperate streams and its variation with stream catchment area. Can. J. Fish. Aquat. Sci. 53, 99e105. Wang, L., Robertson, D.M., Garrison, P.J., 2007. Linkages between nutrients and assemblages of macroinvertebrates and fish in wadeable Streams: Implication to nutrient criteria development. Environ. Manag. 39, 194e212. Wilson, C.B., Walker, W.W., 1989. Development of Lake assessment methods based upon the aquatic ecoregion concept. Lake Reserv. Manag. 5, 11e22. Yuan, L.L., Thomas, D.A., Paul, J.F., Paul, M.J., Kenney, M.A., 2010. Using Stressor-response Relationships to Derive Numeric Nutrient Criteria. United States Environmental Protection Agency, Office of Water, Office of Science and Technology, Washington, D.C.