Estimating Water Quality Parameters Using Surface Spectral Reflectance

Hyperspectral sensing at close range has been considered as an effective tool in estimating water quality parameters.  Many narrow bands allow precise quantification by measuring its pronounced spectral characteristics.  As a fundamental research tool for the application of current and future hyperspectral and even multispectral sensing systems, a field spectroradiometer was used to collect reflectance data.  Derivative spectra indicate the rate of change of reflectance with wavelength (dR(λ)/dλ), which is the slope of the reflectance curve at wavelength λ.  Derivative analysis allows one to correlate the shape of reflectance pattern to chlorophyll concentrations.  Derivative analysis has been applied by researchers in studying the spectral characteristics of chlorophyll and suspended sediments in water.  We have found that 1) first order derivative is able to remove pure water effects while the second derivative can remove suspended sediment effects; 2) the first derivative at 690 nm is useful in estimating chlorophyll concentration in the presence of other water constituents; and 3) derivative spectra is an objective tool in isolating the absorption features of phytoplankton.

       

Mapping Water Quality Parameters Using Satellite Imagery

Phytoplankton are floating or drifting single-cell algae that are primarily transported by water motion.  These organisms are found in all estuarine waters and contribute greatly to overall primary production.  Due to the significant role that phytoplankton play in marine habitats, they are used as indicators of health in systems such as estuaries.  Phytoplankton contain chloroplasts, which absorb and use the underwater light to fix carbon in the form of carbohydrate.  Among the chloroplast pigments, chlorophyll a is common to all phytoplankton although two major color phytoplankton groups, green and brown, also contain chlorophylls b and c respectively.  Thus chlorophyll a is an indicator of the abundance of phytoplankton in the water.  The pronounced scattering/absorption features of chlorophyll a are: strong absorption between 400-500 nm (blue) and at 680 nm (red), and reflectance maximums at 550 nm (green) and 700 nm (NIR).

Remote sensing techniques have been applied to measure chlorophyll a by researchers.  There are some algorithms that have been developed in this endeavor.  Among these algorithms, band ratioing has proven to be advantageous because it tends to allow compensation for variations from atmospheric influences.  In addition, the scattering and absorption characteristics of chlorophyll a can be studied when more than one band is used.  A basic principal of using band ratios is to select two spectral bands that are representative of absorption/scattering features of chlorophyll a. 

Several satellite sensing systems were specifically designed for monitoring chlorophyll a in ocean water, such as Coastal Zone Color Scanner (1978-1986) and the Sea-viewing Wide Field of View Sensor (SeaWiFS).  But they are mostly useful for deep ocean (Case I) waters. Landsat TM/ETM+ data are useful for assessing inland surface water and estuarine systems for several reasons.  The data are economical, routinely available, and archived.  Although the spectral resolution of Landsat TM/ETM+ are modest in quantifying chlorophyll a, the spatial resolution and coverage are adequate for monitoring estuaries.  Landsat TM/ETM+ in conjunction with in situ water sampling provides the means to establishing a relationship between satellite derived reflectance values and chlorophyll a concentrations.  The temporal and spatial distribution of chlorophyll a can therefore be mapped. 

             

Modeling Nutrients Dynamics with GIS*

According to US EPA, nutrients, nitrogen and phosphorus, have been considered as one of the top three causes of use impairment in US waters for more than a decade. Excess nutrients lead to significant water quality problems including harmful algal blooms, hypoxia, declines in wildlife and wildlife habitat, and have been linked to increases in human pathogens. One of my research interests has been modeling nutrients output in a watershed using GIS.   SWAT (Soil and Water Assessment Tool), developed by the US Department of Agriculture-Agricultural Research Service (USDA-ARS), is a physical based, distributed parameter, and time continuous hydrologic simulation model that operates on a daily time step.  The model was developed to nutrient productions, surface runoff, and sediment yields.  The framework of SWAT is illustrated as follows:

The major input variables of SWAT are:

   

The examples of SWAT output:

   

* Special thanks go to Gang Wang, a graduate research assistant, for helping in SWAT modeling