Predicting the water quality of shallow Arctic ponds using remote sensing

Gabriela Tarin, University of Texas at El Paso


Barrow, Alaska is in a region dominated by Arctic tundra of which a substantial part is covered by lakes and ponds. Despite their dominance in the landscape, freshwater ecosystems in the Arctic have been insufficiently studied. It is clear that furthering understanding of how Arctic water bodies are responding to warming will depend on the analysis of changes in the concentration of organic and inorganic constituents in the water; however, scientists are faced with the task of sampling many remote sites in a relatively hostile environment. Thus, the exploration and incorporation of remote methods for monitoring changes in water quality. However, ponds are often excluded from remote sensing studies due their shallow depth and their small size, leading to difficulty in selecting a platform suitable for their small spatial area, and generally shallow depth. The objective of this study was to examine the utility of established optical remote sensing indices collected from both ground-based (JAZ spectrometer) and satellite-based (WorldView-2) measurements of open water reflectance for predicting water quality of arctic tundra ponds. Multiple strong relationships were found between environmental parameters and ground-based reflectance from the JAZ spectrometer. Ground based reflectance appeared to be a strong predictor for measurements of chlorophyll, total suspended solids (TSS) and dissolved carbon compounds. Some of the most useful indices appeared to be the single wavelengths at 682 nm (Index 18) and 806 nm (Index 16), as well as the multiple wavelength ratios 710+820/740 (Index 6), 710+820/675+740 (Index 8) and 700/400 (Index 5). We found that ponds with different characteristics had unique reflectance signatures that could, at least in part, be associated with the differing water chemistries of these different pond types. High concentration of dissolved carbon compounds, especially dissolved organic carbon (DOC) and C440 , in thermokarst ponds in particular, led to very unique signatures and should be examined further to strengthen the utility of established relationships. We found substantially more significant relationships with environmental parameters using the ground-based reflectance data than when we used satellite-based data. Only two ratios proved useful for explaining water quality under both platforms. The ratio of 700:400 was strongly associated with measures of chlorophyll, which are commonly associated with the 700 nm wavelength, as well as measures of carbon (DOC, C440), which commonly peak at 400 nm. The ratio 400/480 was also useful on both platforms for estimating TSS and the spectral ratio. Using a preliminary model for predicting CO2 based solely on remotely-sensed DOC levels, we were able to predict CO2 with a reasonable degree of accuracy in small tundra ponds. These types of models do not exist for shallow freshwater ecosystems, and hold promise for complementing landscape-level estimates of carbon flux following further analyses

Subject Area

Climate Change|Environmental science|Remote sensing

Recommended Citation

Tarin, Gabriela, "Predicting the water quality of shallow Arctic ponds using remote sensing" (2016). ETD Collection for University of Texas, El Paso. AAI10250316.