Abstract
Harmful algae blooms (HABs) can negatively impact water quality, lake aesthetics, and can harm human and animal health. However, monitoring for HABs is rare in Minnesota. Detecting blooms which can vary spatially and may only be present briefly is challenging, so expanding monitoring in Minnesota would require the use of new and cost efficient technologies. Unmanned aerial vehicles (UAVs) were used for bloom mapping using RGB and near-infrared imagery. Real time monitoring was conducted in Bass Lake, in Faribault County, MN using trail cameras. Time series forecasting was conducted with high frequency chlorophyll-a data from a water quality sonde. Normalized Difference Vegetation Index (NDVI) was generally well correlated to chlorophyll-a measured by a sonde (R2 = 0.678 for all data from 5 flights, between 0.323-0.986 for individual flights), while Visible Water Residence Index (VWRI) showed a weaker and less consistent correlation with chlorophyll-a (R2 = 0.027 for all data from 5 flights, between 0.17-0.866 for individual flights). While RGB cameras (trail cameras or UAVs) were useful for visual inspection and spotting blooms, these results suggest that quantitative remote sensing of chlorophyll in Minnesota Lakes should use near-infrared cameras at a minimum. Univariate time series forecasts using sonde chlorophyll-a data were compared using classical (ARIMA, wavelet-ARIMA) and machine learning techniques (LSTM, wavelet-LSTM). Chlorophyll-a was positively correlated to temperature and precipitation, while negatively correlated to conductivity and turbidity. Peak summer chlorophyll concentrations also appeared to be positively correlated to recent precipitation totals. 10-day chlorophyll-a forecasts using univariate LSTM and ARIMA outperformed a multivariate forecast (using conductivity, turbidity, temperature, and precipitation as predictors), suggesting that lower cost monitoring setups (a single chlorophyll probe) may be practical. To assist in understanding meteorological factors impacting interannual variability of blooms in Bass Lake, the relationship between peak summer chlorophyll-a (from Sentinel-2 satellite imagery) and temperature and precipitation were analyzed at Bass Lake. The impact of meteorological factors on patterns in chlorophyll-a for lakes in the Western Corn Belt Plains (WCBP) was also examined, using Sentinel-2 imagery (imagery was available for 160 lakes in the WCBP during 2019 and 2020). Peak summer Chlorophyll-a (from Sentinel-2 imagery) at Bass Lake was positively correlated to 2-week precipitation totals, suggesting a potential role of precipitation induced nutrient loading in initiating blooms; a negative correlation between peak chlorophyll-a and 60-day precipitation totals also suggested that increased residence time during drier periods may be a driving factor as well. While a slight negative correlation between precipitation and peak summer chlorophyll-a was present in a larger scale analysis of 160 WCBP lakes, too many confounding factors were present to show the impact of precipitation on blooms at a broader scale in Minnesota.
Advisor
Guarionex Salivia
Committee Member
Bryce Hoppie
Committee Member
Fei Yuan
Date of Degree
2021
Language
english
Document Type
Thesis
Degree
Master of Science (MS)
Department
Computer Information Science
College
Science, Engineering and Technology
Recommended Citation
Von Korff, B. (2021). Assessing and forecasting chlorophyll abundances in Minnesota lakes using remote sensing and statistical approaches [Master’s thesis, Minnesota State University, Mankato]. Cornerstone: A Collection of Scholarly and Creative Works for Minnesota State University, Mankato. https://cornerstone.lib.mnsu.edu/etds/1161/
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Rights Statement
