Date of Award


Document Type

Honors Thesis (Colby Access Only)


Colby College. Geology Dept.


Jennifer D. Shosa

Second Advisor

Dr. Bruce F. Rueger

Third Advisor

Dr. D. Whitney King


Understanding the flow dynamics of a watershed is important to investigations of a watershed. Numerical modeling is often useful to build a realistic representation of the hydrogeology and to study these flow dynamics in the watershed. Temporal changes in the hydrology make it necessary to define "hydrologic seasons" that separate the year into distinct time periods of similar hydrologic conditions. The hydrologic seasons allow for the seasonal variation in flow dynamics to be expressed in the numerical model. The hydrologic seasons were determined and a 3-D numerical model was built for the Belgrade Lakes Watershed. The Belgrade Lakes Watershed is a small watershed located in south-central Maine. Using 25-years of historical climatological and stream flow data from 1975-1999, eight parameters were used to define seasons' starting and ending dates: (l) temperature, (2) precipitation, (3) evapotranspiration, (4) snowfall, (5) snow cover, (6) ice out, (7) stream discharge, and (8) recharge-runoff balance. From the parameters, the watershed was determined to have six distinct hydrologic seasons. The seasons are: (1) December 18 February IS, (2) February 16 -March 16, (3) March 17 -April 10, (4) April 11 -June 9, (5) June 10 -September 6, and (6) September 7 -December 17. The variable constant flux recharge and runoff rates were determined for each season. A 3-D numerical model of the watershed was built in MODFLOW. To create the model, the hydrogeologic boundaries and the hydraulic parameters were defined. The major boundaries assigned in the model were the watershed divide, bedrock, and the ground surface. Additionally, lakes, streams, eskers, and wetlands were defined in the model before it was run. The model output of the watershed was compared to an interpolated water table from wells in the watershed region. The output was three times higher than the interpolated surface predicted. The variable constant flux recharge and runoff rates determined for each hydrologic season were not applied to the model due to the unrealistic output. The challenges with creating a model that converged are most likely due 10 the lack of data available for the Belgrade Lakes Watershed.


Full-text download restricted to Colby College campus only.


hydrologic seasons, belgrade lakes, watershed, climate