|
Titel |
Estimation Of Chlorophyll-A Concentration Using Three-Band Algorithm On Worldview-2 Images |
VerfasserIn |
Muhittin Karaman, Z. Damla Uça Avcı, Murat Budakoğlu, Serpil Geredeli (Yılmaz), Mustafa Kumral |
Konferenz |
EGU General Assembly 2014
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 16 (2014) |
Datensatznummer |
250096421
|
Publikation (Nr.) |
EGU/EGU2014-11926.pdf |
|
|
|
Zusammenfassung |
The objective of the study is performing an estimation of chlorophyll-a (chl-a)
concentration for an inland turbid salty lake. The study area is selected as Acigol
(Turkey). To retrieve chl-a concentrations, a field study was conducted on July
09th, 2013 to collect water samples at 18 points. Concentrations over the area were
between 0.87-8.72 mg/m3. Due to the average chl-a content which was 6.27 mg/m3,
the lake is categorized as mezotrophic. As satellite data, a high resolution image
acquired by WorldView-2 was used. The data consists of 8 bands that were defined
as water bands according to the EM ranges sensitive to water. Radiometric and
ATCOR atmospheric corrections were executed as preprocessing. Then, for each
sampling point in the image, mean reflectance values in 1*1, 3*3, 5*5, 7*7, 9*9,
11*11, 13*13, 15*15, 17*17, 19*19, 21*21, 51*51 neighborhoods were calculated.
As the method, three-band model was applied to the data in all neighborhoods,
in order to determine the highest correlations between spectral values and chl-a
content. Three-band reflectance model, which was first developed for estimating
pigment contents in terrestrial vegetation, could also be evaluated as a method to
assess chl-a in turbid productive waters as mentioned in the literature. Three-band
model was tested with variable optical properties and it was shown that λ1 should be
around 670 nm, λ2 around 710 nm, and λ3 around 750 nm. Although correlations
were investigated for all neighborhing windows, results were mainly evaluated
for 51*51. Using three-band model for 51*51, λ1= 546 nm (green), λ2= 608 nm
(red), λ3= 659 nm (yellow) wavelengths found to give correlation as 0.9494. The
results of this study show that a strong linear relationship is found between the
chl-a concentration and remotely sensed spectral data, and three-band model is an
effective way to detect the correlations between spectra and chlorophyll-a content. |
|
|
|
|
|