Comparison of spatial interpolation methods for distribution map an unmanned surface vehicle data for chlorophyll-a monitoring in the stream

Abstract

This study aims to comparatively evaluate the interpolation methods to estimate the chlorophyll-a concentrations in a spatial distribution map using the unmanned surface vehicle (USV). We targeted 0.0231 km 2 of the N stream, installed and operated a Chlorophyll-a sensor on a USV, and collected a significant amount of data. These data were extracted via data pre-processing using the data exploration and linear interpolation methods and subsequently utilized for spatial distribution analysis. The inverse distance weighted (IDW) and Kriging interpolation methods were compared to express the chlorophyll-a in a spatial distribution map. With the IDW method, R 2 was 0.81, and the root mean square error (RMSE) was 4.45, and with the Kriging method, R 2 was 0.61, and the RMSE was 6.39. We selected IDW as the optimal interpolation method and manually classified the chlorophyll-a concentrations based on the Korean lake water-quality standard. The analysis of the chlorophyll-a distribution indicated the following: Type Ia (chlorophyll-a 5 μ g / L or less) was 2.6%; Type Ib (chlorophyll-a 9 μ g / L or less) was 9.9%; Type II (chlorophyll-a 14 μ g / L or less) was 19.1%; Type III (chlorophyll-a 20 μ g /L or less) was 27.9%; Type IV (chlorophyll-a 35 μ g /L or less) was 33.2%; and Type V (chlorophyll-a 70 μ g /L or less) was 7.4%. The chlorophyll-a concentrations determined using the USV were successfully represented in the spatial distribution map. • The unmanned surface vehicle (USV) was conducted to monitor the chlorophyll-a concentration distribution in river. • The inverse distance weighted (IDW) was applied for the spatial interpolation method. • The spatial distribution map exhibited considerable potential for the commercialization of remote monitoring technology development using USV.