Scholarly record

ASSESSING DIFFERENT WATER QUALITY PARAMETERS IN THE RIVER SAVA USING AIRBORNE HYPERSPECTRAL REMOTE SENSING TECHNIQUES

Mak Kisevic, Roko Andricevic

Abstract

The majority of current water quality measurements are based on the traditional in-situ point sampling and laboratory analysis of the samples. This method, besides being time and financially more consuming than the remote sensing, also very often lacks a sufficient spatial and temporal resolution to adequately analyse and describe different phenomena (i.e. eutrophication). In this study, we have assessed the possibility to use air-borne hyperspectral remote sensing for mapping concentrations of different water quality parameters in the River Sava. The water quality prediction maps of chlorophyll a (Chl-a), total suspended solids (TSS) and turbidity were made from the hyperspectral data obtained by AISA Eagle airborne instrument using the algorithms developed in the previous study. The 11 measurements taken from the ship platform time and space coincidental with the airborne image were used for the validation of these maps. The results show that the Chl-a model slightly overestimates measured values (NMRS = 14.55%), but generally shows good accuracy with the moderate determination coefficient (R2 = 0.33). TSS and turbidity models display similar accuracy with turbidity values having the highest determination coefficient of all three models (R2 = 0.41). TSS and turbidity prediction maps significantly overlap which is to be expected due to their strong correlation (r = 0.64, p < 0.05).As well, the statistically significant correlations were empirically found of the band ratio R460/R719 with total nitrogen (Ntot, r = -0.72, N = 14) and the band ratio R623/R664 with oxygen values (r = 0.85, N = 14).While these results have to be taken with caution and confirmed with more measurements, they represent a promising first step in the initiative to develop a methodology for the water quality monitoring of the River Sava using remotely sensed data originating from various airborne multispectral and hyperspectral sensors.

Publication Impact Profile

• Captures
• Mendeley - Readers: 4

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2017/31/s12.006

Publication details

Back to publication list