Title: Vitense, K., M.A. Hanson, B.R. Herwig, K.D. Zimmer, J. Fieberg (2018). Data supporting "Predicting total phosphorus levels as indicators for shallow lake management" Authors: Kelsey Vitense, University of Minnesota, viten003@umn.edu Mark A. Hanson, Minnesota Department of Natural Resources Brian R. Herwig, Minnesota Department of Natural Resources Kyle D. Zimmer, University of St. Thomas John Fieberg, University of Minnesota Description: This repository contains data supporting the results reported in Vitense et al. (in revision). From the Methods section of Vitense et al. (in revision): The Minnesota Department of Natural Resources (MNDNR) sampled 118 lakes once in July during each of three consecutive years, 2009-2011. Six lakes were sampled in only two years, and three lakes were sampled just once. TP and fish populations were sampled in each year, and maximum depth was measured either once (9 lakes) or yearly (109 lakes). Lake maximum depth was determined by measuring depths along parallel transects spaced throughout the open water zone of each site. Water samples for TP were collected at two stations in each lake-year and frozen until analysis with persulfate digestion and ascorbic acid colorimetry. TP values are averages for each lake-year. TP values were discretized into two TP classes using the median tip-down TP threshold estimate (50 μg/L) from Vitense et al. (2018): (1) “Stable-Clear” (TP ≤ 50 μg/L, where only the clear stable state exists); and (2) “Bistable-Turbid” (TP > 50 μg/L, where the turbid state exists either in conjunction with the clear state or alone). Fish presence-absence, relative abundance, and community composition were determined using a combination of gears deployed overnight. Three mini-fyke nets targeting minnows and juvenile fishes (6.5-mm bar mesh with 4 hoops, 1 throat, 7.62 m lead, and a 0.69 X 0.99 m rectangular frame opening into the trap) were set overnight in the littoral zone of each lake. One experimental gill net targeting larger-bodied fish species (2 m by 61 m multifilament net with 19.0, 25.0, 32.0, 38.0, and 51.0-mm bar meshes) was deployed along the deepest depth contour in lakes less than 2 m deep or along a 2 m contour in lakes with sufficient depth. All fish sampled were sorted by species and weighed in bulk. Data were summarized as presence-absence and catch per unit effort (CPUE) of fish species belonging to each major feeding guild (planktivores, benthivores, and piscivores), where CPUE was the total biomass collected in the four nets. Several watershed-scale variables were compiled for each lake. These metrics (described below) did not vary between years for each lake (i.e., there is one observation per lake). Lakes were scored on a 0-4 scale according to the degree to which they were connected to downstream and upstream fish sources, ranging from complete isolation (scored as a 0) to connections to seasonal wetlands (scored as a 1), semi-permanent or permanent wetlands (scored as a 2), lakes (scored as a 3), and streams or rivers (scored as a 4). We also evaluated whether the lake or near-lake area contained a dock, residential home or farmstead, subsurface storm sewer, or adjacent road. Details on methods for deriving connectivity metrics are described in Herwig et al. (2010). Lake basin area was determined based on 2008 color air photos (https://www.fsa.usda.gov/programs-and-services/aerial-photography/imagery-programs/naip-imagery/index) used to guide on-screen digitization of lake polygons using ArcGIS software and included both open water and emergent vegetation portions of each lake. Upstream watershed area of each study lake was obtained from and followed the methods of the Minnesota Department of Natural Resources statewide Lake Watershed Project (http://www.dnr.state.mn.us/watersheds/lakeshed_project.html). We derived land cover data in the upstream watershed of each lake by summarizing manually-delineated cover type polygons that were created using on-screen digitizing procedures in ArcGIS. We used 2008 color air photos as the primary interpretive reference for distinguishing cover types, with the 2001 National Land Cover Database (Homer et al., 2007) used to corroborate air photo interpretations as needed. We used the Ecological Classification System (ECS) for Minnesota developed by the MNDNR and the U.S. Forest Service to assign to each lake the ecological province in which it is located, which reflects distinct climatic conditions, vegetation, and biomes (http://www.dnr.state.mn.us/ecs/index.html). We compiled soil texture data in the upstream watershed of each lake in ArcGIS using a map (Miller and White, 1998) of the sand, silt, clay, and organic matter fractions in both the top layer (5cm) of soil and the average fractions in the top 10 layers of soil (1.5m total depth), weighted by layer thickness. We summarized these sand, silt, and clay fractions for each watershed using a weighted average by area of the soil texture map units contained within the watershed. Only four lakes had an organic matter fraction greater than 10%, and we therefore excluded the organic matter fraction from the analysis. We also used a map describing the quaternary geology of Minnesota (Hobbs and Goebel, 1982) to compute the proportion of each upstream watershed containing glacial outwash and gravelly terraces in ArcGIS. Files: 1. TP_DATA.csv = Data used in the analysis. Column descriptions below. a. Lake = unique lake identifier b. Year = year in-lake data was collected (2009, 2010, or 2011) c. TPug = Average TP measured for each lake-year (micrograms/liter) d. TP_discrete = Discretized TP values for classification (StableClear = (TP ≤ 50 μg/L), BistableTurbid = (TP > 50 μg/L) e. PercentWoodland = Proportion of the upstream watershed land cover that is woodlands f. PercentDist = Proportion of the upstream watershed land cover that is disturbed (crop/livestock agriculture + human development) g. undisturbed = Total area of disturbance (crop/livestock agriculture + human development) in upstream watershed (log m^2) h. PercentAg = Proportion of the upstream watershed land cover that is agriculture (crops + livestock) i. lnAg = Total area of agriculture (crops + livestock) in upstream watershed (log m^2) j. CLAY_L1 = Clay fraction in top 5 cm of soil in upstream watershed (weighted average by polygon area) (%) k. CLAY_1.5m = Clay fraction in top 1.5 m of soil in upstream watershed (weighted average by polygon area) (%) l. SILT_L1 = Silt fraction in top 5 cm of soil in upstream watershed (weighted average by polygon area) (%) m. SILT_1.5m = Silt fraction in top 1.5 m of soil in upstream watershed (weighted average by polygon area) (%) n. SAND_L1 = Sand fraction in top 5 cm of soil in upstream watershed (weighted average by polygon area) (%) o. SAND_1.5m = Sand fraction in top 1.5 m of soil in upstream watershed (weighted average by polygon area) (%) p. prop.outwash = Proportion of upstream watershed that is glacial outwash or terraces q. Province = Ecological province. Coded for random forest analysis as follows: 1=Laurentian Mixed Forest Province, 2=Eastern Broadleaf Forest Province, 3=Prairie Parkland Province r. BenthPres = Benthivorous fish present (1=Present, 0=Not present) s. PlankPres = Planktivorous fish present (1=Present, 0=Not present) t. PiscPres = Piscivorous fish present (1=Present, 0=Not present) u. Benth = Catch per unit effort (total biomass in all 4 nets) of benthivorous fish (kg) v. Plank = Catch per unit effort (total biomass in all 4 nets) of planktivorous fish (kg) w. Pisc = Catch per unit effort (total biomass in all 4 nets) of piscivorous fish (kg) x. UpFishSource5 = Degree to which the lake is connected to an upstream fish source: 0= completely isolated 1= connected to seasonal wetland (type 1-3 wetland) 2= connected to semi-permanent or permanent wetland (Type 4 or 5 wetland) 3= connected to a lake 4= connected to a stream/river y. DownFishSource5 = Degree to which the lake is connected to a downstream fish source: 0= completely isolated 1= connected to seasonal wetland (type 1-3 wetland) 2= connected to semi-permanent or permanent wetland (Type 4 or 5 wetland) 3= connected to a lake 4= connected to a stream/river z. RiparianResident = Residence/farmstead in immediate riparian area (1=TRUE, 0=FALSE) aa. StormSewer = Connected to subsurface storm sewer (1=TRUE, 0=FALSE) ab. DockPresent = Dock present on lake (1=TRUE, 0=FALSE) ac. AdjacentRoad = Road nearby lake (1=TRUE, 0=FALSE) ad. Depth = Maximum lake depth (m) ae. BasinArea = Area of lake basin (m^2) af. WAtoLA = Ratio of watershed to lake area License/Restriction Info: These data are protected under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 United States license. References: Herwig, B.R., Zimmer, K.D., Hanson, M. A., Konsti, M.L., Younk, J. A., Wright, R.W., Vaughn, S.R., Haustein, M.D., 2010. Factors influencing fish distributions in shallow lakes in prairie and prairie-parkland regions of Minnesota, USA. Wetlands 30, 609–619. Hobbs, H.C., Goebel, J.E., 1982. S-01 Geologic map of Minnesota, Quaternary geology. Minnesota Geological Survey. Retrieved from the University of Minnesota Digital Conservancy, Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., Wickham, J., 2007. Completion of the 2001 National Land Cover Database for the Conterminous United States. Photogramm. Eng. Remote Sens. 73, 337–341. Miller, D.A., White, R.A., 1998. A Conterminous United States Multilayer Soil Characteristics Dataset for Regional Climate and Hydrology Modeling. Earth Interact. 2, 1–26. Vitense, K., Hanson, M.A., Herwig, B.R., Zimmer, K.D., Fieberg, J., 2018. Uncovering state-dependent relationships in shallow lakes using Bayesian latent variable regression. Ecol. Appl. 28, 309–322. Vitense, K., Hanson, M.A., Herwig, B.R., Zimmer, K.D., Fieberg, J., in revision. Predicting total phosphorus levels as indicators for shallow lake management. Ecol. Indic.