In most north temperate lakes such as Lake Mendota, Wisconsin, phosphorus (P) often limits primary productivity. The largest source of P to the lake is the watershed. When P loads to the lake are large, blue-green algae (cyano-bacteria) dominate and form nuisance blooms. My overall research goals were to identify linkages between factors that occur at the landscape scale with factors that occur at the scale of algae. My specific research objectives were to: (1) develop a spatially explicit model to quantify the effect of land use change on annual P loading; (2) compare the relative contribution of internal and external P sources to Lake Mendota's epilimnetic P-budget during summer; and (3) determine the predictability of blue-green algal blooms at the daily scale in Lake Mendota.

In collaboration with others, I developed a simple model to estimate nonpoint-source P loading from land to water using geographic information system (GIS) databases. The model has only three parameters and accounts for the distribution of land uses with distance from open water. The greatest contribution to loading was found to come from a heterogeneous riparian corridor that surrounds the lake and streams. The width of this corridor varies among years depending on annual precipitation. This concept of a "variable source area" has rarely been applied to P loading in large watersheds, and yet may provide a useful framework for examining the importance of landscape features on characteristics of surface waters.

Because internal sources of P in lakes can also be large and variable, I compared internal and external loads to the epilimnion of Lake Mendota during two summers. Internal loading in deep, stratified lakes occurs through two major mechanisms: hypolimnetic entrainment of P across the thermocline and epilimnetic sediment release. Although loading from entrainment has been suggested to be large, it can be difficult to characterize. I developed a method to measure the input of P that occurs by entrainment using continuous temperature profiles measured with thermistors attached to a buoy and a datalogger. Internal loading from hypolimnetic entrainment made up a large part of the P budget during both years, and was about 8 times larger than external loading during an average loading year.

However, because algae respond to environmental fluctuations at time lags ranging from 1 - 15 days, I examined the response of blue-green algae to internal and external P supply and environmental variables at the daily scale in Lake Mendota during summer 1993. Surface blooms occurred every 20 - 30 days, and lasted from 1 - 3 days. My results show that blue-green algal dynamics are not predictable at the daily scale. In addition, water column chlorophyll concentrations and surface bloom occurrence were not strongly related. This is important because water column chlorophyll is commonly measured and predicted using water quality models, but it is surface blooms that are perceived by the public and need to be managed.