Landsat 8: Providing continuity and increased precision for measuring multi-decadal time series of total suspended matter

Abstract

Abstract The water clarity of many inland water bodies is under threat due to intensifying land use pressures in conjunction with changes in water levels that result from increasing demand and climate variability. The recent launch of Landsat 8 coupled with Geoscience Australia's recent reprocessing of the Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM +) archives over the whole of Australia to a consistent surface reflectance product enables sub continental scale spatio-temporal analysis of freshwater optical water quality in support of monitoring and decision making for water management agencies. In this research, we present an objective assessment of the potential of Landsat 5 TM, Landsat 7 ETM + and Landsat 8 Operational Land Imager (OLI) data for monitoring inland water quality dynamics over a number of lakes and reservoirs with a range of optical water types in New South Wales and Queensland, Australia. We used bio-optical modelling to develop sensor-specific total suspended matter (TSM) retrieval algorithms that account for the difference in relative spectral response between Landsat 7 ETM + and Landsat 8 OLI. We were able to compare the suitability of the different sensors for optical water quality measurements using water bodies that fell within Landsat path overlaps where Landsat images of surface reflectance were acquired within 24 h between Landsat 5 TM and Landsat 7 ETM + or Landsat 7 ETM + and Landsat 8 OLI. These water bodies represent a range of hydrological and limnological conditions, and enabled us to assess: 1) the comparability of TSM measurements retrieved from each sensor, and 2) the surface reflectance to image noise characteristics of Landsat 7 ETM + and Landsat 8 OLI. Comparisons of lake surface reflectance and noise equivalent reflectance difference show that the improved radiometric resolution and increased quantization of Landsat 8 OLI relative to Landsat 7 ETM + significantly reduce image noise and spectral heterogeneity, indicating that Landsat 8 OLI data are expected to provide more precise water quality retrievals relative to Landsat 7 ETM +. We found that: 1) the TSM retrievals from the different sensors are highly comparable; 2) Landsat 5 TM overestimated TSM relative to Landsat 7 ETM + by 6.4%; and 3) Landsat 7 ETM + overestimated TSM relative to Landsat 8 OLI by only 1.4%. Retrieved TSM values were highly correlated with independent in situ data acquired within 24 h of satellite overpass ( r = 0.99) with a mean average error of 14 mg/L. The results demonstrate that time series analysis of TSM retrievals can be conducted across a wide range of lakes at the sub-continental scale to characterise the multi-decadal TSM dynamics.