RIASSUNTO
ABSTRACT
This paper introduces an integrated system comprising a UAV mounted with a spectroradiometer, which is capable of collecting reflectance and parameters of water quality simultaneously. Reflectance and parameters of water quality, such as chlorophyll-a and turbidity were collected in five estuaries along the Jiangsu Coast, China. Using a power function with these sensitive wavebands, glint can be removed from measured reflectance spectra. This study shows how a UAV-based spectrometer system is used to measure water reflectance and retrieve water quality parameters accurately.
INTRODUCTION
Estuaries are coastal water bodies which have a great ecological importance to the environment. They serve as habitat and nurseries to numerous unique organisms like birds and shellfish species. These wetlands can filter runoff in these buffer zones to stabilize the coastal shoreline and protect the coastal areas, inland habitats, and human communities from floods, and storm surges from hurricanes.
These vital ecological attributes are highly dependent on the overall ecosystem's health, and water quality is a major determinant of an estuary's ability to support healthy food webs. The water quality management is therefore an important issue concerning this critical ecosystem.
Aquatic ecosystems like estuaries have been polluted by nutrients over the last few decades on a worldwide scale, due to the growing population and the necessity to extend provision of developed land which leads to an alteration of natural pathways for runoff dispersal. The monitoring and management of these water bodies are therefore essential.
Traditionally, water quality, monitoring is implemented by point-wise measurements, in-situ, but this approach only offers data for specific sampling points rather than the overall water body (Dornhofer et al., 2016). Moreover, multispectral satellite imageries have been widely applied to water quality monitoring (Gitelson et al., 2008; Teodoro et al., 2007; Zhang et al., 2010). However, water quality retrieval models, obtained from remotely sensed images and in-situ observations, lack of robustness and transferability. One of the main reasons is that the spectral reflectance, derived from the remotely sensed images, are not synchronized with ground observations. Such limitation restricts its application in routinely water quality monitoring and environmental management. Hence a platform without atmospheric effects and offering more control over environmental variables during data acquisition is valuable for studying water-light interactions and further water constituent modelling.