Relationships between Residence Time and Cyanobacterial Blooms in a Nutrient-Rich River System Michael F. Coveney, John C. Hendrickson, Erich R. Marzolf, Rolland S. Fulton, Jian J. Di, Clifford P. Neubauer, Dean R. Dobberfuhl, Greeneville B. Hall St. Johns River Water Management District, Palatka, FL, USA Hans W. Paerl University of North Carolina Chapel Hill, Institute of Marine Sciences, Morehead City, NC, USA Edward J. Phlips University of Florida, Gainesville, FL, USA
Residence Time and Cyanobacterial Blooms in the St. Johns River Outline 1. Existing impairment in water quality due to nutrients/algal blooms 2. Relationships between algal blooms, nutrients, and hydrology 3. Thresholds for adverse ecological effects of algal blooms 4. Would water withdrawals exacerbate the adverse effects of algal blooms?
Water Bodies in the St. Johns River Basin Impaired by Excess Nutrients Water bodies with established total maximum daily loads (TMDLs) or listed as impaired by the Florida Dept of Environmental Protection St. Johns River ~ 500 km Impaired Palatka & Racy Point Lake George
Chlorophyll-a Concentration 1995-2005 100 95% N S 90% 80 75% Median Chl-a g L-1 60 40 25% 10% 5% Algal Bloom 20 0 Fulton Pt. Mandarin Pt. Shands Br. N Racy Pt. Palatka L. George S L. Monroe L. Harney L. Poinsett
St. Johns River at Mandarin, August 2005, Microcystis
Color Pt-Co Units Relationships Water Chemistry Water Age (April-Oct, log Y-axes, 1995-2005) 0.32 1024 0.16 512 LG Pal Racy TP mg L -1 0.08 0.04 0.02 LG Pal Racy 0 50 100 150 200 250 256 128 64 32 16 0 50 100 150 200 250 Water Age d Water Age d TN mg L -1 3.2 1.6 LG Pal Racy Chl-a µg L -1 256 128 64 32 16 LG Pal Racy 8 0.8 0 50 100 150 200 250 4 0 50 100 150 200 250 Water Age d Water Age d
Chl-a vs TP Relationship Slope Depends on Water Age Chl-a ug L -1 128 64 32 16 LG Pal Racy Water Age 80 d Chl-a ug L -1 256 128 64 32 16 LG Pal Racy Water Age > 80 d 8 0.00 0.05 0.10 0.15 0.20 TP mg L -1 Negative relationship between Chl-a and TP at short retention times (LG & Pal p = 0.05, Racy NS) 8 0.00 0.05 0.10 0.15 0.20 TP mg L -1 Positive relationship between Chl-a and TP at long retention times (All regressions p < 0.05) Monthly mean values, Apr Oct, 1995 2005; log Y-scales; regressions ln(y) = mx+b
Hydro-Ecological Models Predict algal bloom metrics (dependent variables) Bloom magnitude (Chl-a) Dinoflagellate abundance Use hydrologic prediction (independent) variables: Water age and variables derived from water age Multiple linear and logistic regression Data sets 11 yr (1995 2005) Bloom duration N 2 -Fixation
Example of Ecological Metric - Algal Bloom Duration Abundance of zooplankton, e.g. cladocera, is reduced during extended algal blooms 200 Cladocera Ind L -1 150 100 50 85%tile LG Pal Racy 0 0 50 100 150 200 Days Continuous Algal Bloom
Cladocera Ind L -1 Decline in Zooplankton Seasonal or Algal Bloom Effect? 400 300 200 100 Cladocera p = 0.033 (neg) No Bloom Long Bloom Copepods Ind L -1 250 200 150 100 50 Copepods NS No Bloom Long Bloom Nauplii Ind L -1 800 600 400 200 0 Jan Mar May Jul Sep Nov Nauplii Month 14 NS 12 p = 0.001 No Bloom Long Bloom Rotifers 10 3 Ind L -1 10 0 Jan Mar May Jul Sep Nov 8 6 4 2 Rotifers (pos) Month No Bloom Long Bloom 0 Jan Mar May Jul Sep Nov Month 0 Jan Mar May Jul Sep Nov Month
Hydro-Ecological Models Dependent Variables: Algal Bloom Metrics Algal Bloom Metric Effect(s) Measured Variable Duration of freshwater algal blooms Altered zooplankton community; reduction in fish production Duration of longest annual bloom Magnitude of freshwater algal blooms Change ( ) in N load Marine algal blooms 1) Altered phytoplankton community; cyanobacterial toxins 2) Depletion of dissolved oxygen; effects on fish reproduction, growth, and mortality Additional N loading Potential toxic species Maximum annual bloom chl-a Annual mass N added via N 2 -fixation Maximum annual dinoflagellate biovolume
Prediction (Independent) Variables Based on Water Age (Residence Time) 1) Water age for five quarterly periods starting with the last quarter of the previous year, plus two growth season periods A Oct-Dec 2) Include mean, maximum, and minimum water age for each period B Jan-Mar April-Oct April-Aug C Apr-Jun D Jul-Sep E Oct-Dec 3) Include the inverse of each water age. Inverse water age is positively related to flow but without negative values at low-flows Monthly mean values Racy Pt & SJSR16
Hydro-Ecological Models multiple linear and logistic regression 8 regression models needed to predict 4 algal bloom metrics across multiple river segments Used 7 linear regression models and 1 logistic regression model Linear regression models 3 to 7 independent variables Adjusted R 2 values of 0.80 to 0.97
Example Prediction of Maximal Bloom Chl-a Observed Maximal Chl-a µg L -1 Observed by Predicted Maximal Bloom Chl-a Multiple Linear Regression Adjusted R 2 = 0.88 Variable Std Regr Coeff MinAgeD -2.318 invminaged -1.764 MeanAgeD 1.367 MaxAgeE 1.297 invmeanagee 0.755 invmean_age_apr_oct 0.636 invmaxagea 0.540 Predicted Maximal Chl-a µg L -1 Data set: Segments 3 & 4 LG12 & Racy Pt
Conclusions - 1 The St. Johns River is impaired by cyanobacterial blooms caused by high nutrient levels Blooms are summertime events and are exacerbated by low-discharge Blooms cause altered food webs, low DO, algal toxins, and increased N loading through N 2 - fixation
Conclusions-2 Relationships between algal blooms and nutrients depend strongly on hydrology At low river discharge: Lower phosphorus, dissolved color, and turbulence Increased chl-a (relaxed light limitation) Algal bloom metrics (e.g. magnitude, duration) are predictable from residence time (water age) with regression models Modeled water withdrawals (~10 6 m 3 d -1 *) caused negligible worsening in algal blooms * 262 x 10 6 gal d -1
Any Questions? The St. Johns River Water Supply Impact Study Final Report: Chapter 8, Plankton http://floridaswater.com/watersupplyimpactstudy/