A comparison between professionally (Florida Department of Environmental Protection) and volunteer (Florida LAKEWATCH) collected trophic state chemistry data in Florida

Abstract

Florida Department of Environmental Protection (FDEP) professional biologists and Florida LAKEWATCH volunteers sampled 27 Florida lakes simultaneously to measure the concentrations of total phosphorus, total nitrogen, and chlorophyll. Each program used their Standard Operating Procedures for both field and laboratory activities to determine data comparability. Results showed that LAKEWATCH data were nearly equivalent to FDEP's, which were collected using stringent quality assurance (QA) protocols and analyzed in a National Environmental Laboratory Accreditation Conference (NELAC)-certified laboratory in compliance with the state's QA rule. The R² values for paired comparisons of total phosphorus, total nitrogen, and chlorophyll were 0.97, 0.90, and 0.97, respectively, indicating functionally equal values between programs. Properly trained volunteers should be viewed as partners, and volunteer monitoring can be embraced as a robust tool for obtaining credible, cost-effective data. Studies like this demonstrate that data collected by volunteer monitoring programs are suitable for regulatory decisions, if such programs and agencies work together to ensure necessary data quality and documentation.

Key words:

• chlorophyll
• citizen science
• lake monitoring
• methods comparison
• nitrogen
• phosphorus
• quality control

Two centuries ago almost all scientists made their living in some other profession (Silverton 2009). For example, Charles Darwin (1809–1888) sailed on the Beagle as an unpaid companion to Captain Robert FitzRoy, not as a hired naturalist (Silverton 2009). Through time, citizens have participated in studies of their natural surroundings out of curiosity and a sense of stewardship for their land and water. Many of these individuals collect scientific data without formal academic training or financial compensation. The United States has nearly 900 volunteer-monitoring initiatives that conduct scientific investigations in conjunction with governmental agencies, generally as a professional public–private partnership (EPA 1998). The objectives for such volunteer monitoring activities range from educational outreach to the generation of data used for regulatory purposes (Loperfido et al. 2010, Mackechnie et al. 2011). Quality assurance and documentation requirements for these data vary, depending on the ultimate use of the information.

Florida LAKEWATCH is a volunteer water quality monitoring program developed by the University of Florida in 1986 with a primary goal of collecting scientifically credible total phosphorus, total nitrogen, chlorophyll a, and water clarity data from a large number of Florida lakes. To ensure the credibility of the data, LAKEWATCH staff conducted extensive quality assurance and quality control activities. Previous studies have shown that data collected by LAKEWATCH volunteers was strongly correlated (r > 0.99) and comparable to those collected by professionals (Canfield et al. 2002). One recurring concern about the credibility of the Florida LAKEWATCH data is that the data collected by volunteers and professionals in the comparability study conducted by LAKEWATCH (Canfield et al. 2002) were all generated in a single laboratory (LAKEWATCH laboratory) with the same chemist and technicians. Additionally, the LAKEWATCH laboratory does not have National Environmental Laboratory Accreditation Conference (NELAC) accreditation. To alleviate these recurring concerns, Florida LAKEWATCH and the Florida Department of Environmental Protection (FDEP) conducted another independent comparability study.

The FDEP is the regulatory agency charged with protecting Florida's water and air resources. In carrying out this charge, the Florida Legislature instructed FDEP to establish, by rule, appropriate quality assurance requirements for environmental data submitted to the department to ensure the scientific validity and legal defensibility for all data used for regulatory purposes (Chapter 403.0623, Florida Statutes). Subsequently, FDEP established and fully complied with the Quality Assurance (QA) Rule (Chapter 62-160, Florida Administrative Code). This rule requires samplers to follow Standard Operating Procedures (SOPs) for sample collection and preservation and requires labs to be certified by NELAC. LAKEWATCH procedures involving sample preservation, holding times, and laboratory analytical methods were thoroughly tested (Canfield et al. 2002). The methods are inconsistent with requirements of the QA Rule, however, and therefore there are potential issues concerning use of the LAKEWATCH data in the regulatory arena.

Because FDEP assessments would benefit from including LAKEWATCH data, this study was conducted to determine the comparability of water chemistry results between the 2 entities. Total phosphorus, total nitrogen, and chlorophyll a samples were simultaneously collected and analyzed from a range of Florida lakes using both Florida LAKEWATCH and standard FDEP field and laboratory methods.

Methods

Between September and November 2011, FDEP biologists and LAKEWATCH volunteers sampled 27 Florida lakes across a gradient of nutrient concentrations. FDEP and LAKEWATCH collected surface grab water samples from the same lake on the same day, with FDEP obtaining 3 replicate samples from one station and LAKEWATCH samplers sampling 3 stations spatially on each lake. FDEP personnel collected samples for total phosphorus, total Kjeldahl nitrogen (TKN), nitrite+nitrate nitrogen (NO₂-N+NO₃-N), and chlorophyll a (both uncorrected and corrected for phaeophytin; hereafter chlorophyll) concentrations. LAKEWATCH volunteers collected samples for total phosphorus, total nitrogen, and chlorophyll a (uncorrected for phaeophytin). FDEP's TKN and NO₂-N+NO₃-N were added together for comparison to LAKEWATCH's total nitrogen (TN) values.

LAKEWATCH field collections and laboratory analyses followed the methods of Canfield et al. (2002). Volunteers collected surface water samples for nutrient analyses in 250 mL acid-washed plastic Nalgene bottles and filtered water on sites or at their homes to collect planktonic algae for chlorophyll analysis. Both nutrient samples and chlorophyll filters were then frozen for later transportation to the University of Florida's LAKEWATCH laboratory where samples were analyzed. All samples for this study were analyzed between 1 and 2 months after collection, which is within the prescribed LAKEWATCH holding time of 1 to 6 months. For total phosphorus determination, LAKEWATCH personnel used a persulfate digestion method from Menzel and Corwin (1965) and measured concentrations with the colorimetric procedures of Murphy and Riley (1962). For total nitrogen determination, LAKEWATCH personnel followed a method described in Bachmann and Canfield (1996), in which the sample was prepared using an autoclave and persulfate digestion (all nitrogen oxidized), and then all NO₃ in the sample was measured and reported as total nitrogen. For chlorophyll analysis, laboratory technicians extracted pigment by soaking filters in 90% ethanol, heated to 78 C for 5 minutes (Sartory and Grobbelaar 1984). They then used the trichromatic equation (Method 10200 H, APHA 2005) to calculate the concentration of chlorophyll. LAKEWATCH reported total phosphorus and chlorophyll to 1 μg/L and total nitrogen to 10 μg/L.

FDEP field methods followed FDEP field SOPs (http://www.dep.state.fl.us/water/sas/sop/sops.htm). Field personnel acidified surface water samples for nutrient analysis to a pH <2, and laboratory technicians analyzed the samples within 28 d. Water samples for chlorophyll analyses were kept on ice for transport to the laboratory, where technicians filtered the chlorophyll sample within 48 h of sample collection, froze the filters, and analyzed them within a 28 d holding limit. All samples were cooled to <6 C between field collection and delivery to the laboratory.

FDEP laboratory methods followed FDEP internal lab SOPs (http://www.dep.state.fl.us/labs/library/lab_sops.htm). For total phosphorus, FDEP used EPA Method 365.1 (Revision 2.0), the ascorbic acid method with absorbance reading at 880 nm. The samples were first analyzed using the Seal AQ2 Discrete Analyzer. Samples with phosphorus concentrations below 50 μg/L were then analyzed using the Bran+Luebbe Segmented Flow Analyzer, and the detection limit was 4 μg/L. For NO₂-N+NO₃-N, FDEP used EPA method 353.2 (Revision 2.0), the cadmium reduction method. The analyses were run on a Lachat Quickchem 8500 Autoanalyzer, and the detection limit was 4 μg/L. For TKN, FDEP used EPA method 351.2 (Revision 2.0), the block digestion and salicylayte method. The analyses were run on a Seal AQ2 Discrete Analyzer, and the detection limit was 80 μg/L. For chlorophyll analysis, FDEP used Standard Methods 10200H (APHA 2005), in which the filter was macerated with a tissue grinder and steeped in 90% acetone to extract chlorophyll from algal cells. Uncorrected chlorophyll was calculated using a trichromatic equation. For chlorophyll, FDEP reported results with 2 significant figures, and those results were used in the current comparison.

Table 1 Summary statistics (total phosphorus [TP], total nitrogen [TN] and chlorophyll uncorrected for phaeophytin [CHL]) for comparability study of 27 Florida lakes sampled between September and November 2011 by Florida LAKEWATCH (LW) and Florida Department of Environmental Protection (FDEP). FDEP total nitrogen values were the sum of Kjeldahl Nitrogen and nitrite+nitrate nitrogen. Relative percent difference (RPD) calculated as the absolute value of 100*(FDEP – LW)/(FDEP+LW/2).

Parameter and Source	Mean	Coefficient of Variation (%)	Minimum	Maximum
FDEP TP (μg/L)	20	69	4	51
LW TP (μg/L)	18	74	3	56
RPD TP (%)	15	—	0	33
FDEP TN (μg/L)	995	70	95	3100
LW TN (μg/L)	980	69	250	3010
RPD TN (%)	10	—	0	90
FDEP CHL (μg/L)	18	140	1	120
LW CHL (μg/L)	19	135	1	121
RPD CHL (%)	13	—	0	61

Table 2 Regression and paired t-Test results comparing total phosphorus (TP), total nitrogen (TN) and chlorophyll uncorrected for phaeophytin (CHL) concentrations from 27 Florida lakes sampled between September and November 2011 by Florida LAKEWATCH (LW) and Florida Department of Environmental Protection (FDEP). All data were log₁₀ transformed to meet assumptions of normality.

	Regression Results FDEP (Data) = b0 + b1*LW (Data)
Statistics	TP (μg/L)	TN (μg/L)	CHL (μg/L)
Intercept	0.10	−0.29	−0.01
Intercept Different From 0.0	yes	no	no
Slope	0.96	1.1	0.99
Slope Different From 1.0	no	no	no
R^2	0.97	0.90	0.97
	Paired t-Test Results
Log Mean Difference (FDEP − LW)	0.05	−0.01	−0.03
Probability of > |t|	0.0001	0.69	0.11
Geometric Mean Difference (FDEP − LW)	1.13	−0.98	−0.94

FDEP collected 3 replicate samples at one station to examine sampling/analysis precision, while LAKEWATCH volunteers only collected one sample at that corresponding station and one sample at 2 other stations to examine spatial variance in their lake. All data were log₁₀ transformed to meet the assumptions of parametric analyses (Sokal and Rohlf 1981). For all paired comparisons and regression analyses, the FDEP's replicate sample with the earliest collection time was compared to the one LAKEWATCH sample collected from the corresponding station. Linear regression analysis and 2-tailed paired t-tests were used to compare results from the sampling teams for each of the 3 parameters. To determine if the slopes of the paired comparisons were significantly different from 1, dummy data representing a 1 to 1 relation (27 point incorporating the range of original data) were created and added to the dataset for each variable. A coding variable (CV) was added setting original data as 0 and the dummy data as 1. Then an analysis of covariance was conducted with the following model:

If b3 was significant in the model then the slope of the original data was significantly different from 1.

All statistics were conducted using JMP software (SAS 2000) using a significance level of p < 0.05. The relative percent difference (RPD) between FDEP and LAKEWATCH data was calculated as the difference between the 2 values divided by the average of the 2 values.

Results and discussion

The 27 lakes used in this comparison study ranged from oligotrophic to eutrophic, encompassing a wide range of Florida lake types (Table 1). The distribution of data for total phosphorus, total nitrogen, and chlorophyll was similar between FDEP and LAKEWATCH results (Table 1). The mean RPD values between FDEP and LAKEWATCH data for total phosphorus, total nitrogen, and chlorophyll were 15%, 10%, and 13%, respectively (Table 1).

The coefficient of determination (R²) for the regression that compared total phosphorus concentrations measured by FDEP and LAKEWATCH was 0.97, demonstrating strong comparability of results reported by the 2 organizations (Fig. 1a; Table 2). The slope of the regression was not significantly different from 1, and the intercept was only slightly different from 0. The paired t-test statistics also showed a significant difference between the 2 sampling groups, with FDEP measuring slightly higher values. The geometric mean difference, however, was only 1.12 μg/L. For 16 of the 19 lakes with total phosphorus greater than the practical quantitation limit (PQL; FDEP results), the RPD was less than or equal to the acceptable lab duplicate limit of 20% (Table 1; FDEP 2008), and the remaining 3 had RPD values of 26, 30, and 32%. The 8 lakes for which the FDEP results were less than the PQL had RPD values from 0 to 33%. We expected the RPD to be higher for values near the method detection limit.

Figure 1 Relations between data, including (a) total phosphorus, TP; (b) total nitrogen, TN, and (c) chlorophyll, CHL concentrations collected and analyzed by Florida Department of Environmental Protection (FDEP) personnel and data collected and analyzed by Florida LAKEWATCH (LW). Data are from 27 Florida lakes sampled between September and December 2011. The linear regression line is solid and the dashed line represents the 1:1 line.

The coefficient of determination (R²) for the regression between total nitrogen concentrations measured by FDEP and LAKEWATCH was 0.90 (Fig. 1b; Table 2). The slope of the regression was not significantly different from 1 and the intercept was not significantly different from 0, indicating that the calculated FDEP total nitrogen concentrations are comparable with the values reported by LAKEWATCH. Paired t-test statistics also showed no significant difference between the 2 sampling organizations, with FDEP's geometric mean only 0.98 μg/L less than LAKEWATCH data. For 24 of the 27 lakes, the RPD was less than or equal to the acceptable lab duplicate limit of 20%, and RPD values for the remaining 3 lakes were 21, 36, and 90% (Table 1).

The coefficient of determination (R²) for the regression between uncorrected chlorophyll concentrations measured by FDEP and LAKEWATCH was 0.97 (Fig. 1c; Table 2). The slope of the regression was not significantly different from 1 and the intercept was not significantly different from 0, indicating high comparability between datasets. Paired t-test statistics also confirmed no significant difference between the 2 sampling organizations, with LAKEWATCH's geometric mean chlorophyll only 0.94 μg/L less than FDEP results. For 24 of the 27 lakes, the RPD was less than or equal to the acceptable lab duplicate limit of 25% and the RPD values for the 3 remaining lakes were 29, 41, and 61% (Table 1; DEP Lab SOP BB-29-1.4).

This comparison study demonstrates that the total phosphorus, total nitrogen, and chlorophyll data provided by LAKEWATCH are nearly equivalent to data generated by FDEP's NELAC-accredited laboratory. The only statistically significant difference observed was for total phosphorus, where the geometric mean of FDEP data averaged 1.12 μg/L higher than the geometric mean of LAKEWATCH data. Although it was statistically significant, this small difference would not likely affect assessments to determine compliance with numeric nutrient standards in Chapter 62-302, F.A.C. (water quality standards) or in time series analyses to detect total phosphorus trends over time.

Florida has more than 7700 lakes. With an increasing need to monitor these lakes during a period of decreasing budgets, it is advantageous to utilize data collected by volunteer groups such as LAKEWATCH. Although probabilistic sampling approaches, such as those employed by USEPA for the National Lakes Survey, are useful for statistically estimating the status of lakes statewide, citizens are often mostly concerned about “my lake” (Canfield and Bachmann 2009). LAKEWATCH can help provide the information residents request on lakes of specific interest to them.

This study demonstrates that FDEP can use the 25+ years of LAKEWATCH data on several hundred lakes and multiple estuaries to assess trends in nutrient and chlorophyll concentrations and be confident that any trend would be based on data of known and acceptable quality. If decreasing water quality is detected in any aquatic systems (early warning method or the “LAKEWATCH Canary”) then FDEP can focus limited resources with their NELAC sampling protocols to gather data required for regulation that may face legal challenges. Other volunteer programs around the country should consider conducting comparison studies to show agencies charged with regulating aquatic systems the value of their volunteer monitoring data.

To use LAKEWATCH data for regulatory proceedings, however, LAKEWATCH would have to make several changes in their operating procedures following the QA Rule. For example, LAKEWATCH would have to correct chlorophyll for phaeophytin; moderately increase their current amount of laboratory quality control data involving calibration, standards, blanks, duplicates, and spikes; and begin tagging data with qualifier codes (as described in the QA Rule). Some items in the QA Rule, however, would be difficult if not impossible to accomplish with a volunteer monitoring program including collecting field blanks with deionized water. Each change in the program would increase the time and cost of processing samples and therefore decrease the cost effectiveness of a volunteer monitoring program. FDEP, in partnership with LAKEWATCH, is developing program-specific quality assurance procedures that take into account logistical limitations associated with a volunteer monitoring program but still provide the documentation to ensure that the LAKEWATCH water quality data are of sufficient quality for regulatory decision-making. This study clearly demonstrates that properly trained citizen scientists are capable of collecting high quality data and that these volunteers should be viewed as valuable partners in Florida's effort to protect and manage water resources.

Acknowledgments

We thank Jerry Brooks for initiating this cooperative study and all the LAKEWATCH volunteers who adjusted their schedules to participate in this study.