USING AN UNMANNED AERIAL VEHICLE EQUIPPED WITH A MULTISPECTRAL SENSOR TO OPTICALLY MAP WATER QUALITY PARAMETERS IN DRINKING WATER RESERVOIRS

Surface drinking water infrastructure is an integral part of the development and sustainability of societies around the world. However, these surface freshwater resources have been a challenge to monitor due to the vast number of in-situ samples needed to accurately quantify constituents, expenses of equipment, coordination of personnel, and lab cost. Lake Bloomington and Evergreen Lake (Central Illinois) are two vital surface water features that serve as the drinking water reservoirs for the City of Bloomington. Both reservoirs receive high nitrate and sediment inputs from streams with >90% row crop agriculture within the watershed.

We utilized an Unmanned Aerial Vehicle (UAV) coupled with a five-band multispectral image sensor to monitor turbidity and algae, two important drinking water parameters, in the reservoirs. By using the UAV, along with in-situ data collected the same day as the flight, we aim to answer the following questions: 1.) What are the challenges of remote sensing over a homogeneous setting (such as a lake), and 2.) Is it possible to detect change in the water quality at the surface of the lake using one or more spectral image analyses?

Preliminary results show that the UAV imagery can produce a 6-cm pixel size with greater than 80% coverage at each sample site and regularly above 90% coverage when upscaled to 3-m resolution. An initial transformation of both resolutions to NDVI (Normalized Difference Vegetation Index) produces spatial patterns for algae within the reservoirs and highlights differences between the reservoirs. With the NDVI images, we were able to visually identify clumped algal mats consisting of spirogyra. Higher resolution images were able to identify individual floating mats (baseball sized), which can often be overlooked in regular color imagery. The UAV shows potential for regularly monitoring surface water quality of reservoirs even in what would be considered adverse weather conditions for comparable imaging done by satellites (Sentinel and Landsat).