Mapping Submerged Aquatic Vegetation Using RapidEye Satellite Data: The Example of Lake Kummerow (Germany)

Abstract

Submersed aquatic vegetation (SAV) is sensitive to changes in environmental conditions and plays an important role as a long-term indictor for the trophic state of freshwater lakes. Variations in water level height, nutrient condition, light availability and water temperature affect the growth and species composition of SAV. Detailed information about seasonal variations in littoral bottom coverage are still unknown, although these effects are expected to mask climate change-related long-term changes, as derived by snapshots of standard monitoring methods included in the European Water Framework Directive. Remote sensing offers concepts to map SAV quickly, within large areas, and at short intervals. This study analyses the potential of a semi-empirical method to map littoral bottom coverage by a multi-seasonal approach. Depth-invariant indices were calculated for four Atmospheric & Topographic Correction (ATCOR2) atmospheric corrected RapidEye data sets acquired at Lake Kummerow, Germany, between June and August 2015. RapidEye data evaluation was supported by in situ measurements of the diffuse attenuation coefficient of the water column and bottom reflectance. The processing chain was able to differentiate between SAV and sandy sediment. The successive increase of SAV coverage from June to August was correctly monitored. Comparisons with in situ and Google Earth imagery revealed medium accuracies (kappa coefficient = 0.61, overall accuracy = 72.2%). The analysed time series further revealed how water constituents and temporary surface phenomena such as sun glint or algal blooms influence the identification success of lake bottom substrates. An abundant algal bloom biased the interpretability of shallow water substrate such that a differentiation of sediments and SAV patches failed completely. Despite the documented limitations, mapping of SAV using RapidEye seems possible, even in eutrophic lakes.

Canopy Reflectance Modeling of Aquatic Vegetation for Algorithm Development: Global Sensitivity Analysis

“…At present, the technologies of remote sensing are widely used for mapping and classification of aquatic vegetation [6][7][8]. Another major application of wetland remote sensing is parameter inversion, which includes mainly biomass [9][10][11][12] and leaf area index (LAI) [13,14].…”

“…The variance D (−i) could be obtained just as in Equation (4) while the range of j was [1, ω i /2]. Finally, the first and total order sensitivity index would be obtained by Equation (5) and (6).…”

Seasonal Variation in Spectral Response of Submerged Aquatic Macrophytes: A Case Study at Lake Starnberg (Germany)

“…Heblinski et al [36] documented spatial vegetation dynamics of Lake Sevan, Armenia, with algorithms that can differentiate bottom coverage types as well as several macrophyte species and sediment types. To exclude the influence of the water column and water depth, Roessler et al [14] and Fritz et al [37] used a method of depth-invariant indices to differentiate between bottom substrates. In-situ reflectance spectra of littoral bottom coverage (e.g., macrophytes or sediments) are very helpful to control atmospheric and water column corrections of remote sensing data and to distinguish macrophyte signal from water column attenuations [6,36,38].…”

Indicator Values of Emergent Vegetation in Overgrowing Lakes in Relation to Water and Sediment Chemistry

“…In addition, these studies most often refer to lakes of considerable size, in contrast to small lakes where intense overgrowth processes observed. Many papers refer to the short period of research covering the last decades [45,47]. Limited recognition of the occurrence of submerged vegetation is also the result of high intensity of phytoplankton blooms, variability of water optical parameters, macrophyte coverage and differentiation or similarity of plant phenology [21,46].…”