Skip to main content
ARS Home » Research » Publications at this Location » Publication #187419

Title: APPLICATION OF HYPERSPECTRAL IMAGERY AND DIGITAL VIDEOGRAPHY TO MANAGE ALGAL BLOOMS IN AQUACULTURE SYSTEMS: CURRENT STATUS.

Author
item Thomson, Steven
item Zimba, Paul

Submitted to: American Society for Photogrammetry and Remote Sensing Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 11/4/2005
Publication Date: 3/13/2006
Citation: Thomson, S.J., P.V. Zimba. 2006. Application of hyperspectral imagery and digital videography to manage algal blooms in aquaculture systems: Current status. In: Proceedings of the 20th Biennial Workshop on Aerial Photography, Videography, and High Resolution Digital Imagery for Resourece Assessment, CD-ROM. American Society for Photogrammetry and Remote Sensing Proceedings(ASPRS), Bethesda, MD.

Interpretive Summary: Detection of pigments that are present in species of algae that cause off-flavor in catfish and fish mortality is necessary to provide best management options. Remote sensing can be used to detect these pigments by the reflectance of light at wavelengths specific to these algal species. Many pond constituents (both organic and inorganic) mask detection of these pigments because they also respond to these wavelengths. It is thus difficult to isolate particular wavelengths that are specific to the harmful algae. In previous research, a model was developed that used three bands (650 nm, 710 nm, and 740 nm) in mathematical combination to enhance the specific spectral signature of chlorophyll-a (Chl-a), a major indicator of water quality and the presence of potentially harmful algae. To determine the robustness of the model, the model was tested against pond sampling data from a previous year (2003), which included sampling for Chl-a. The model fit the data from 2003 very well, although single outlying data points were observed for a few ponds. These ponds showed Phaeophytin (Chl-a degradation product) levels up to 22 times higher than the average from the other ponds sampled on the same day and may account for the data outliers. More tests and analyses need to be conducted to confirm possible influence of these and other pond constituents on model response. Concepts for a remote sensing system were outlined that use narrow band filters with band centers at 650 nm, 710 nm, and 740 nm to detect Chl-a in catfish production ponds. This model will be useful in managing catfish production ponds more accurately.

Technical Abstract: Detection of harmful algal blooms in Case II hypereutrophic aquaculture systems continues to be a challenge. Attempts to isolate certain pond constituents have been difficult because both organic matter and suspended sediments can mask detection of these components. A three band reciprocal reflectance model, originally developed to detect Chlorophyll-a (Chl-a) in terrestrial vegetation, had been tuned previously for detection of Chl-a in catfish production ponds at the USDA-ARS Stoneville, MS research facility using pond samples and spectroradiometer data. To test the model’s robustness, this study applied the tuned model to two other data sets from the same ponds. The strength of the model was confirmed by goodness of fit to these data. The good results are also noteworthy by the fact that spectroradiometer readings for the calibration data set were taken slightly below the water surface whereas readings for the evaluation sets were taken above the water surface, the latter possibly introducing specular reflection effects. Some model fits indicated peculiar outliers which, on preliminary analysis, indicated the presence of very high concentrations of Phaeophytin (Chl-a degradation products). Remote sensing systems to detect Chl-a in catfish production ponds can be adapted by selection of narrow-band optical filters with band centers at 650, 710, and 740 nm.