Relationship Between Development and Condition of Lakes in Minnesota's Northern Lakes and Forests Ecoregion

Abstract

Human development in shorelands and watersheds has the potential to impact the condition and sustainability of Minnesota lakes. Shoreland and local zoning ordinances have guided development in recognition of this potential impact, but must be based on scientific evidence of land-lake linkages. The impacts of development must also be distinguished from the effects of lake and watershed characteristics that naturally alter lake condition. The objectives of this project were to: (1) develop classification systems for lakes based on (a) measures of lake condition and (b) measures of shoreland population density, and (2)develop methodologies to assess the cumulative effects of development on lakes by relating indicators of human activity to indicators of lake condition. GIS and statistical analysis techniques were used to relate indicators of lake condition, in particular Secchi transparency, to lake, shoreland and watershed characteristics. Existing lake condition data were obtained from two digital repositories: the Minnesota DNR Fisheries Data Warehouse and EPA's STORET database. Data were analyzed for two time periods, 1977-79 and 1994-96, so as to temporally match lake condition with available GIS land use data. The spatial extent of the study is the Northern Lakes and Forest (NLF) ecoregion, which covers 68,243 km2 in northeastern Minnesota and contains 5,408 lakes >. 10 ha, the majority of which are oligotrophic to mesotrophic. Secchi transparency of 589 lakes sampled in 1994-96 ranged from 0.6 to nearly 12 m. Of the eight clearest lakes, three were abandoned mine pits. Average Secchi transparency was 2.75 m for the 1977-79 time period, but increased to 3.17 m for the 1994-96 time period, a significant increase in lake clarity over time. Highly colored lakes were less transparent, and the correlation between Secchi transparency and lake color (r = -0.480, p = 0.000) was greater than the correlation between Secchi transparency and total P ( r = -0.365, p = 0.014). Multivariate analysis of seven lake condition parameters revealed that 73% of the variance in lake condition was explained by just two lake principal components (LPCs): LPCl was positively related to pH, alkalinity, total dissolved solids, and conductivity, whereas LPC2 was positively related to Secchi transparency and negatively related to chlorophyll A and total P. These principal components were used to classify lakes into four lake condition classes: deep transparency low alkalinity, shallow transparency low alkalinity, shallow transparency high alkalinity, and deep transparency high alkalinity. The area of lakes studied ranged from 5 to 51, 7 48 ha, and maximum lake depth ranged from 1 to 70 m. Average shoreland land cover within a 1000' buffer surrounding each lake was 69% forest and brushland, 11% wetland, 8% water (ie., adjacent lakes), 6% grassland, 5% urban, and less than 1% cultivation and mining. With the exception of mining within the shoreland zone, which was associated with clearer mine pit lakes, there were no significant correlations between individual measures of development and Secchi transparency. Lakes with people living around them were significantly clearer than those without, indicating that people choose to live on clearer lakes. Shoreland population density was grouped into five classes based on the natural breaks method, but the only significant difference in Secchi transparency were between the most populous class (>. 844.4 people/km2), which represented the three lakes with the most extreme urbanization in the ecoregion, and the two classes with intermediate population densities (37.5-361.1 people/lm12). The transparency of the most populous class was not significantly different than the transparency of the least populous class (0-3 7 people/km2). Stepwise multiple regression between Secchi transparency and original environmental variables revealed that maximum lake depth was the single variable that consistently had the greatest influence on Secchi transparency, the deeper the lake the greater the clarity. Maximum lake depth provided 47 to 75% of the explanatory power of the stepwise multiple regressions developed using original variables, much more than any variable related to development. This means that deep lakes are naturally less sensitive to cumulative impacts than are shallow lakes. Stepwise multiple regression between Secchi transparency and principal components derived from environmental variables revealed that the following natural conditions were associated with clearer lakes: watersheds with loamy vadose soils, watersheds with moderately acid soils of very low erodibility, and lakes with other lakes in their shoreland zone. Lakes with more wetlands in their shoreland and watershed were less transparent. Lakes with urban shorelands were less transparent, but lakes in watersheds with private land ownership were clearer than lakes in watersheds with public non-wilderness ownership. Seasonal home development was associated with clearer lakes, but lakes with the most seasonal home development tended to be deep, which may have overridden any negative effects of shoreland development. Cultivated crops in the shoreland were associated with circumneutral soils, which were associated with reduced clarity. Overall, there was little evidence that development was detrimental to Secchi transparency except at the highest levels of population density and urbanization.