New protocols for evaluation of Optical Water Types (OWTs) of lakes and wetlands using multispectral EO datasets

Abstract:

Lakes and wetlands are important surface water bodies and are important constituents of the terrestrial water cycle. They play a major role in maintaining the water-security in many regions, and therefore, assume socio-economic significance. However, they are under tremendous pressure all over the world due to various natural and anthropogenic factors, and therefore, require an immediate attention for their conservation. The available studies on wetland have given much less importance to the internal dynamics of the lakes and wetlands, which is primarily driven by hydrology and Land Use Land Cover (LULC) changes. Here, we propose to use the Optical Water Types (OWTs) concept to understand the hydrodynamics within the lakes and wetland.

The OWTs are the aquatic counterpart of terrestrial LULC classification. Traditionally, the OWTs have been obtained using hyperspectral datasets. However, since lake/wetland hydrodynamics assessment require multi-temporal datasets, the data availability and computation requirements hinder the applicability of hyperspectral-based OWTs. These shortcomings can be managed by using multispectral Earth Observation (EO) datasets such as Landsat and Sentinel-2 series imageries which are freely available and have high temporal resolution. Accordingly, a protocol has been devised to obtain the OWTs using indices such as Normalized Difference Water Index (NDWI), Normalized Difference Vegetation Index (NDVI), and Normalized Difference Turbidity Index (NDTI) derived from the multispectral EO dataset. These indices range between -1 to +1 and are known to capture the remote sensing reflectance, absorption, and scattering behavior of the optically sensitive constituents of water such as chlorophyll, turbidity, suspended organic and inorganic matter. The negative values of indices were discarded, and the positive values were classified in 4 classes at equal intervals of 0.25 and a unique ID was given to each interval. A collection of the IDs of different indices for a given pixel defines its OWT. Since the intervals are pre-defined, the resulting OWTs across space and time represent the same underlying physical phenomenon. This protocol was applied at various lakes and wetlands of India in different hydrogeomorphic settings such as Chilika (coastal lagoon), Kaabar Tal (a floodplain wetland), Osman Sagar, Hussain Sagar, and Himayat Sagar (lakes in Indian Cratonic region). The OWTs were further correlated with various in-situ water quality parameters of some of these lakes and wetlands to provide a broader scope of the OWTs in restoration and management activities. The protocol developed in this research are useful to distinguish the coastal water and turbidity dominated OWTs in the Chilika lake as well as suspended matter dominated OWTs in the Kaabar wetland. The Osman Sagar and Himayat Sagar have very little turbidity and mostly dominated by the freshwater inlets. The Hussain Sagar is dominated by the municipal sewage water and has higher dissolved organic matter. Our results in different hydrogeomorphic settings exhibit the universal applicability of the protocol developed for derivation of OWT.