Chlorophyll-a Estimation Research Articles

Remote chlorophyll-a estimates for inland waters based on a cluster-based classification

Accurate estimates of chlorophyll-a concentration (Chl-a) from remotely sensed data for inland waters are challenging due to their optical complexity. In this study, a framework of Chl-a estimation is established for optically complex inland waters based on combination of water optical classification and two semi-empirical algorithms. Three spectrally distinct water types (Type I to Type III) are first identified using a clustering method performed on remote sensing reflectance (Rrs) from datasets containing 231 samples from Lake Taihu, Lake Chaohu, Lake Dianchi, and Three Gorges Reservoir. The classification criteria for each optical water type are subsequently defined for MERIS images based on the spectral characteristics of the three water types. The criteria cluster every Rrs spectrum into one of the three water types by comparing the values from band 7 (central band: 665nm), band 8 (central band: 681.25nm), and band 9 (central band: 708.75nm) of MERIS images. Based on the water classification, the type-specific three-band algorithms (TBA) and type-specific advanced three-band algorithm (ATBA) are developed for each water type using the same datasets. By pre-classifying, errors are decreased for the two algorithms, with the mean absolute percent error (MAPE) of TBA decreasing from 36.5% to 23% for the calibration datasets, and from 40% to 28% for ATBA. The accuracy of the two algorithms for validation data indicates that optical classification eliminates the need to adjust the optimal locations of the three bands or to re-parameterize to estimate Chl-a for other waters. The classification criteria and the type-specific ATBA are additionally validated by two MERIS images. The framework of first classifying optical water types based on reflectance characteristics and subsequently developing type-specific algorithms for different water types is a valid scheme for reducing errors in Chl-a estimation for optically complex inland waters.

Water quality monitoring in Basque coastal areas using local chlorophyll-<inline-formula><math display="inline" overflow="scroll"><mrow><mi>a</mi></mrow></math></inline-formula> algorithm and MERIS images

Accurate estimation of chlorophyll-a (chl-a), a proxy of the eutrophication risk, is necessary in coastal areas for the assessment of water quality in accordance with European Direc- tives. Local parameterization of remote sensing algorithms is useful to cope with the variability and specificity of optically-active in-water constituents. Using the Bay of Biscay coastal waters, affected by Basque river runoffs, as a case study, the objectives of this investigation are to: 1. develop an empirical algorithm to estimate water surface chl-a for the optically-complex Basque coastal waters; 2. explore the influence of suspended matter, phytoplankton species, and pigment content on the algorithm developed for medium resolution imaging spectrometer instrument (MERIS) imagery; 3. compare the local algorithm to three ocean color algorithms (OC4v6, Gitelson's algorithm, and the OC5); and 4. apply the local algorithm to the MERIS images. For this purpose, two surveys were undertaken within the study area, the Batel-1 survey in 2007, and the Batel-2, in 2009. The empirical algorithm was developed with remote sensing reflectances (Rrs), undertaken with a TriOS field spectrometer, and chl-a measured in situ from the Batel-2 survey. The algorithm was not affected by different concentrations of suspended matter in surface waters, within the range from 0.0 to 6.6 g · m −3 . There was no significant effect of 23 accessory pigments found in the area on the algorithm. Eighty-four Rrs and chl-a measure- ments from the Batel-1 survey were used to validate the local algorithm and to compare it with output of the other algorithms. The local algorithm provided the lowest root-mean-square differ- ence (RMS ¼ 1.7 mg · m −3 ), the best correlation with the observed data (R ¼ 0.8), together with the best slope-intercept combination between predicted and observed chl-a (slope ¼ 0.5, intercept ¼ 0.6). The chl-a algorithm developed here for MERIS imagery can assist in the assess- ment of water ecological status in the southeastern part of the Bay of Biscay, in a cost-effective manner. © 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). (DOI: 10.1117/ 1.JRS.6.063519)

고해상도 위성영상을 이용한 낙동강 유역의 클로로필-a 농도 추정

본 연구에서는 Moderate Imaging Spectroradiometer(MODIS), Sea-viewing Wide Fieldfo-view Sensor(SeaWiFS), Medium Resolution Imaging Spectrometer(MERIS) 등의 광역관측 위성 영상을 이용한 해수나 연안수의 클로로필 농도 분석을 통해 가능성이 확인되었던 밴드 비를 이용한 비교적 간단한 추정 모델을 수체의 크기와 폭이 현저히 작고 탁도가 있는 하천에 대해 클로로필- a농도값을 추정하고자 고해상도 위성영상에 Two-band 및 Three-band reflectance 모델을 적용하여 가능성을 파악하였다. 특히 RapidEye 영상을 이용하여 일반적으로 탁도가 있는 수체에 대해 Red와 NIR 영역을 활용하는 이들 모델에 Red-edge(RE) 밴드를 적용하였다. Red와 NIR을 이용한 Two-band Reflectance 모델은 계산식의 결정계수 R2 값이 0.38로 유의성 없는 결과를 나타내었다. 그러나 RapidEye의 Red-edge (RE) 파장대를 이용한 Red-RE Two-band 모델과 Red-RE-NIR Three-band 모델을 이용한 계산식에 대해서는, 2차함수에 의한 Three-band 모델의 결과는 Red-RE Two-band 모델의 결과와 통계적인 값이 거의 유사하였고 Two-band와 3차함수에 의한 Three-band 모델 추정식은 각각 0.66, 0.73 의 R2값을 나타내어 Red-edge 밴드의 적용 가능성을 보였고, 실측치와의 Root Mean Square Error (RMSE)는 24.8, 22.4 mg m-3, Relative Percent Difference(RPD)는 각각 1.30, 1.29로 1.5 이하의 대략적인 추정(Approximate Prediction) 수준을 나타내었다. 고해상도 위성영상에 Red-RE-NIR Three-band 모델을 적용한 계산식을 이용해 대략적인 추정이지만 가장 유의한 수준의 클로로필- a농도를 추정할 수 있었다. 영상에서 추정된 클로로필 -a분포를 비교하였을 때 3차함수에 의한 Three-band 모델 추정식이 Two-band 모델에 비해 낮은 값의 분포를 보였다. 향후 하천의 스펙트럼을 실측하여 파장별 부유물질, 유기물과의 상관성 및 클로로필 농도와의 간섭 정도를 시뮬레이션하여 보정식을 산출·적용한다면 탁도가 다소 높은 하천에서의 클로로필 -a농도 계산식의 정확도를 더욱 높일 수 있을 것으로 기대된다.

Remote sensing of chlorophyll-a concentration for drinking water source using genetic algorithms (GA)-partial least square (PLS) modeling

Accurate estimation of phytoplankton chlorophyll-a (Chl-a) concentration in turbid waters through remote sensing is a challenge due to the optical complexity of water constituents. Reflectance spectra and concurrent water quality parameters of 225 samples across the Shitoukoumen Reservoir, the drinking water resource for Changchun City, were used to retrieve Chl-a concentration with high total suspended matter (TSM) during 2006–2008. A combination of genetic algorithms and partial least square (GA-PLS) model was established for Chl-a retrieval through GA to select sensitive spectral variables and PLS for regression. To compare GA-PLS performances, the widely accepted three-band algorithm was implemented for Chl-a concentration estimation. Both GA-PLS and the three-band algorithm have stable performance for the aggregated dataset (R2=0.85 and 0.81; RPD=3.95 and 3.61; relative RMSE=31.7% and 34.2%), with the GA-PLS model performing marginally better. The temporal transferability of the models was validated with the dataset collected in 2006 and 2007 respectively as independent dataset, showing that GA-PLS outperformed the three-band algorithm. Our result also indicated that relative error [(Chl-apredicted−Chl-ameasured)/Chl-ameasured] showed good linear relation to TSM: Chl-a ratio (R2=0.84), which implied that TSM concentration exerted significant impact on the accuracy of Chl-a estimation in this case study. As the results were derived from a large number of samples representing a wide range of spatiotemporal variations of pigment under TSM (3.7–472.8mg/L) concentration influence, the GA-PLS model has great potential for Chl-a estimation for inland waters with similar backgrounds. Nevertheless, the three-band algorithm also has its own merit considering its simplicity for implementation.

Algorithms for remote estimation of chlorophyll-a in coastal and inland waters using red and near infrared bands

Remote sensing algorithms that use red and NIR bands for the estimation of chlorophyll-a concentration [Chl] can be more effective in inland and coastal waters than algorithms that use blue and green bands. We tested such two-band and three-band red-NIR algorithms using comprehensive synthetic data sets of reflectance spectra and inherent optical properties related to various water parameters and a very consistent in situ data set from several lakes in Nebraska, USA. The two-band algorithms tested with MERIS bands were Rrs(708)/Rrs(665) and Rrs(753)/Rrs(665). The three-band algorithm with MERIS bands was in the form R3=[Rrs(-1)(665)-Rrs(-1)(708)]×Rrs(753). It is shown that the relationships of both Rrs(708)/Rrs(665) and R3 with [Chl] do not depend much on the absorption by CDOM and non-algal particles, or the backscattering properties of water constituents, and can be defined in terms of water absorption coefficients at the respective bands as well as the phytoplankton specific absorption coefficient at 665 nm. The relationship of the latter with [Chl] was established for [Chl]>1 mg/m3 and then further used to develop algorithms which showed a very good match with field data and should not require regional tuning.

Medium resolution imaging spectrometer (MERIS) estimation of chlorophyll-a concentration in the turbid sediment-laden waters of the Changjiang (Yangtze) Estuary

Satellite estimation of chlorophyll-a (Chl-a) concentration in the Changjiang Estuary is challenging for ocean-colour retrieval algorithms. The Changjiang Estuary is characterized by suspended-sediment-dominated waters in the mouth and optically complex case 2 waters offshore of the mouth. Satellite ocean-colour products show that high-sediment loads of estuarine waters can cause overestimations or invalid flags of Chl-a concentration. A synthetic chlorophyll index (SCI) was designed for extracting chlorophyll information and for minimizing the influence of sediments on the remote-sensing reflectance spectrum. An SCI algorithm, a quadratic polynomial function of the SCI versus Chl-a concentration, was applied to the estimation of Chl-a concentration from Medium Resolution Imaging Spectrometer (MERIS) images. The overestimation of Chl-a concentration was corrected. The SCI algorithm has applications for MERIS estimation of Chl-a concentration in turbid waters with a moderate to high suspended-sediment concentration.

Equivalent Leave Water Reflectance Based on In Situ Spectral Measurements In the Taihu Lake

Leave water reflectance is an important parameter in the study of water optical characteristics.To better interpret the effect of cyanophytes contamination on water optical characteristics,the authors conducted in situ measurement of spectral reflectance and water sampling in the Taihu Lake on 10 and 11,November 2008.Remarkable effects were observed in leave water reflectance of the cyanobacterial water,leading to an obvious absorption peak in the red region and an increase in the near-infrared region.Equivalent leave water reflectance based on FY-3A and MODIS band settings was derived by using the spectral response functions.Furthermore,the authors used the Ration Index(RI) model for the estimation of chlorophyll-a on 12,November 2008,and observed high determination coefficients R2=0.72,which were further used to map the chlorophyll-a distribution.The results obtained will be helpful to the further evaluation of optical characteristics and water quality.

Testing the spectral decomposition algorithm (SDA) for different phytoplankton species by a simulation based on tank experiments

The spectral decomposition algorithm (SDA), which is a new approach for the simultaneous estimation of chlorophyll-a (Chl-a) and non-phytoplankton suspended solid (NPSS) concentrations in Case 2 waters using satellite data, was proposed by our previous study. Here, we investigated the effect of phytoplankton composition on the SDA based on the tank experiments using cultured samples for five phytoplankton species (three cyanobacteria, one green algae and one diatom) and lake water samples collected from Lake Kasumigaura from September 2004 to August 2005 except for April 2005. The results showed that the SDA-based models obtained from the cultured samples (culture-based models) showed high accuracies for Chl-a and NPSS estimations in all phytoplankton species (root mean square error [RMSE] about 16.2 μg l − 1 for Chl-a and about 11.0 mg l − 1 for NPSS). However, the Landsat Thematic Mapper (TM) band combinations, which gave the smallest RMSE of the Chl-a and NPSS estimations, differed among the species. In addition, the culture-based models could apply to lake water samples with similar accuracies if the optimal standard reflectance spectra (SRS) of end-members in the water body could be determined. This implies the potential of the SDA-based model as follows: (1) it is less site- and time-specific compared with conventional empirical methods (single band, band ratio, and arithmetic band calculation); (2) it can be proposed by a tank experiment or by a simulation using bio-optical modelling.

Simultaneous Estimation of SS and Chlorophyll in Shallow Water through Remote Sensing

The simultaneous estimation of SS and Chlorophyll-a (Chl.a) in shallow water through remote sensing was tried in this study. Lake Teganuma located in Chiba Prefecture was selected as test site. Twelve spectral / Chl.a data sets were acquired using outboard boat in the lake on 28 May 2009 and 22 Oct. 2008. These data sets were related to SS or Chl.a using simple single band or two band ratio models. As result, statically significant correlation was observed between a of the simulated ALOS/AVNIR-2 Band (760-890nm) and SS, or reflectance ratio of the simulated AVNIR-2 band 3 (610-690nm) and band 4 and Chl.a. These models were validated though the ALOS/AVNIR-2 data and the buoy data set on 20 May 2008 and 28 Oct. 2009.

Estimation of Chlorophyll-a and Suspended Solids Concentration in Rich Concentration Water Area with Remote Sensing Technique

Estimation of Chlorophyll-a and Suspended Solids Concentration in Rich Concentration Water Area with Remote Sensing Technique

The neural network model for estimation of chlorophyll-a with water temperature in Lake Taihu

The advantage of neural network method for estimating water quality parameters of complex water body has been approved. Using in-situ measurement data of chlorophyll-a concentration, imageries of MODIS 250m and retrieval model of water temperature, we develop two single-hidden-layer BP neural network models for estimating chlorophyll-a in Lake Taihu: Model NN1 without temperature input and Model NN2 with temperature input. The training method is used by Levenberg-Marquardt algorithm, and the early-stage determinationin the modeling is used to improve generalization. The results show that: the estimation precision of the two models is high, in which the estimation precision of neural network input with temperature has been improved although the test is not significant.

Development of Chlorophyll-$a$ Algorithm for Ocean Colour Monitor Onboard OCEANSAT-2 Satellite

An empirical chlorophyll algorithm has been developed using the coincident in situ chlorophyll-a and remote sensing reflectance Rrs measurements from global ocean waters. The basic data set used for developing the algorithm was obtained by merging the bio-optical data from the global NASA bio-Optical Marine Algorithm Data (NOMAD) (~2438 spectra from ~3000 stations) and from the waters of the northern Arabian Sea (~159 spectra) collected by the Space Applications Centre, Ahmedabad, India. The chlorophyll-a concentration ranged from 0.01 to 50.0 mg ldr m-3 for the data set used. Regression analysis between chlorophyll-a concentration and remote sensing reflectance in different bands and a combination of band ratios was performed. Algorithms using modified cubic polynomial (MCP) regression of Rrs ratios with chlorophyll-a concentration showed good estimates of chlorophyll-a in full range of 0.01 to 50.0 mg ldr m-3 of the merged data set. However, the best results were obtained by using MCP regression between maximum band ratio (MBR) of Rrs (443, 490, 510 nm)/Rrs 555 nm with chlorophyll-a concentration having an r2 of 0.96 and rms error of 0.12 for log-transformed data. The developed MBR-based algorithm named Ocean Colour Monitor (OCM)-2 chlorophyll algorithm was compared with the OC4v4 algorithm routinely used the for Sea-viewing Wide Field-of-view Sensor (SeaWiFS) data processing. For the used data set, OC4v4 algorithm overestimated chlorophyll-a concentration for > 5.0 mg ldr m-3 and yielded an r2 of 0.90 with rms error of 0.23, when compared to the newly developed OCM-2 chlorophyll algorithm. It is proposed to use this OCM-2 chlorophyll algorithm with OCEANSAT-2 OCM data to be launched in the third quarter of the year 2008 by the Indian Space Research Organisation.

衛星データを使った内湾の鉛直クロロフィル推定のためのガウスモデルの適用可能性評価

An objective of this research is to evaluate the feasibility of the Gaussian distribution model's application for the vertical Chlorophyll-a (Chl-a) estimation in inner bay by the satellite remote sensing. The in-situ Chl-a (N=15) observed as data for a basic Gaussian model verification in Hiroshima Bay between 2004 and 2005 was used. Consequently, the vertical Chl-a in the bay is to approximate by the Gaussian model using six parameters. However, if the equation for the estimation of Gaussian model parameter is not adjusted every time, the vertical Chl-a estimation only from surface Chl-a data obtained from the satellite data is difficult. The most difficult Gaussian model parameter to estimate from the surface Chl-a was Cmax (Chl-a maximum) and Zmax (the depth of the Chl-a maximum).

Less is better: Uncorrected versus pheopigment-corrected photometric chlorophyll-a estimation

Photometric determinations of uncorrected and pheophytin-corrected chlorophyll-a concentrations were compared with HPLC-determinations. All determinations were based on the same water samples, which were taken during a two-year survey of seven lakes with varying trophy and morphology. Uncorrected and pheophytin-corrected chlorophyll-a concentrations determined from the same samples differed significantly. Furthermore, pheophytin-corrected chlorophyll-a concentrations differed significantly from the HPLC-determinations, whereas the uncorrected photometric measurements did not differ. In addition, measurements of the same samples with different photometers, showed a significantly increased variation for the pheophytin-corrected chlorophyll-a concentrations. The presented data suggest the abandonment of the acidification of sample extracts resp. the determination of pheophytin-corrected chlorophyll-a concentrations, because i) the methodological procedure for the photometric determination of uncorrected chlorophylla concentrations is less time-consuming, ii) the accuracy of the resulting data is better and iii) measurements obtained with different photometers are better comparable.

Quantitative Estimation of Chlorophyll-a and Total Suspended Matter Concentration with Landsat ETM Based on Field Spectral Features of Lake Taihu

The spectra were measured with hand-spectral instrument in Lake Taihu, and the water samples were collected in situ and analyzed in the lab. The correlations between chlorophyll-a (CHL) and total suspended matter (TSM) and the measured spectra reflectivity were analyzed, on the basis of which, CHL and TSM concentrations were estimated by using the combination of the field bands, respectively. Then by the comparison of the bands of Landsat ETM, the most practicable ones were selected, which would be validated for the estimation of CHL and TSM content by using the Landsat ETM images. Finally, lots of regression models were discussed to educe the estimation models. Two of the most precise ones were used to estimate the CHL and TSM concentrations. The conclusion are as follows: (1) Landsat TM/ETM is one of the appropriate data resources in the multi-spectral remote sensing, but it isnt the most appropriate, (2) it is more precise to estimate CHL content with the third band of ETM by some certain function. In addition, it is also better to estimate CHL content with the combination of ETM3 and ETM1, (3) the correlation between ETM4 and TSM is also very close, and the most precise band to estimate TSM is ETM4/ETM1 , (4) the image window size has an effect on the precision to estimate the water quality parameters, and 5 by 5 or 7 by 7 is more appropriate to estimate CHL content, while 3 by 3 or less is more appropriate to estimate TSM content.

SeaWiFS data analysis and match-ups with in situ chlorophyll concentrations in Danish waters

For the year 1999 all Sea viewing Wide Field of view Sensor (SeaWiFS) scenes of the Danish waters from the North Sea to the Baltic Sea were browsed, and a total of 47 SeaWiFS scenes with reasonably low cloud cover and, therefore, potential in situ match-ups were found and processed. The in situ data used as match-ups were collected on routine monitoring cruises by Danish and Swedish environmental authorities. A few stations in the North Sea, Skagerak and the western Baltic Sea were sampled, while most stations were located in Kattegat and the inner Danish waters. A turbid water SeaWiFS atmospheric correction algorithm was applied, since the standard SeaWiFS algorithm for chlorophyll-a (CHL) has been shown to be fairly inaccurate in turbid coastal waters. This is due to both inaccurate atmospheric and to relatively high and variable abundance of yellow substance. The application of the turbid atmospheric correction substantially improved the SeaWiFS CHL estimates. Regressions between SeaWiFS estimates using the OC2 and OC4 algorithms used in the SeaDAS software (versions 3.3 and 4.0, respectively) and in situ CHL values were made as well, and regression with a number of other possible reflectance ratios with SeaWiFS channels. The best correlation was found to be R2=0.54 using a double-ratio algorithm using both R510/R555 and R443/R670, while the OC4v4 had the second best correlation of R2=0.39. Among other single ratios, the R510/R555 had the highest correlation with CHL, which was expected since this is also the ratio that OC4v4 most often switches to in the waters investigated here. The range of CHL concentrations in this study was rather limited (all but three points from 0.5–3 mg m−3) so there is a need for inclusion of more data to expand the concentration range. This should be possible using also data from 2000, 2001 and onwards and, hereafter, a more ‘stable’ empirical algorithm can be derived for the Danish waters.

Some Notes on Water Color in Keweenaw Bay (Lake Superior)

Spectral subsurface irradiance reflectance—intrinsic water color—was derived from above-water radiance measurements using a hand-held spectroradiometer along a transect on Keweenaw Bay, Lake Superior. The reflectance spectra were typical of oligotrophic lacustrine waters. The reflectance peak wavelength shifted from 484 nm at stations farthest offshore to 540 nm near the head of the bay. This change coincided with a decrease in Secchi-disk depth from 16 to 8 m, and an increase in concentrations of chlorophyll a and total suspended matter from about 0.45 to 0.60 mg m −3 and from 0.3 to 0.5 g m −3, respectively. The concentration of chromophoric dissolved organic matter (gilvin), expressed as the absorption of filtrate at 440 nm, varied between 0.11 and 0.2 m −1. Like almost all inland waters, Keweenaw Bay should be classified as a Case 2 water due to the concentrations of gilvin and inanimate particles relative to plankton biomass. A model for chlorophyll-a estimation from spectral reflectance that adequately predicted concentrations in mesotrophic to highly eutrophic Case 2 waters elsewhere gave negative values when applied to the Keweenaw Bay transect. Evidently, there is a need of algorithm development for oligotrophic lacustrine waters. Estimates improved using a modified blue to green band ratio algorithm previously applied for remote sensing of oceanic waters. In optimization of semi-empirical algorithms for estimation of plankton biomass in Lake Superior, absorption by gilvin as well as by inanimate particles merits special consideration.

Application of empirical neural networks to chlorophyll-a estimation in coastal waters using remote optosensors

This paper presents chlorophyll-a estimation in coastal waters off the Gulf of Finland using remote optosensors. Concurrent remote optosensor data and in situ measurements of water quality were obtained in the study area. Significant correlations were observed between digital values and chlorophyll-a measurements. The results as a case study show that the estimated accuracy of chlorophyll-a retrieval using neural networks is higher than the accuracy of chlorophyll-a estimation using regression analyzes in the area. The study also shows one example why remote optosensors are critical to monitor water quality in coastal areas such as the Gulf of Finland.

Adaptation of an algorithm for chlorophyll-a estimation by optical data in the oligotrophic Gulf of Eilat

Downwelling irradiance and upwelling radiance from the water surface were measured in the Gulf of Eilat (Red Sea) on 55 separate occasions in a pelagic station in the period from January 1994 to December 1996. The reflectance was calculated in seven of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) spectral channels throughout the visible spectrum. Concurrent with the optical measurement, water samples were taken to the laboratory and the concentrations of chlorophyll-a (Chl-a) were determined. The ratio of the reflectance in the blue ( R 443) to the reflectance in the green ( R 550) was regressed against the Chl-a concentration, and showed that the best-fit of the three datasets (1994, 1995 and 1996) assumed an exponential form, as found in previous work in oligotrophic waters. The coefficients determined in the current study were, however, different from those published in the literature; the reason for this probably originates in the composition of water constituents in the studied area. Small cells dominate the phytoplankton in the pelagic waters in the Gulf of Eilat, and the concentration of suspended and soluble materials of terrestrial origin is very small. The suggested algorithm enabled the prediction of Chl-a concentration within 0.08 mg m @ 3 . It was shown that calibration of the coefficients, based on empirical data, is important for increasing its prediction accuracy.

Estimation of Chlorophyll-a for Full Completed and Incompleted Regression Model.

Estimation of Chlorophyll-a for Full Completed and Incompleted Regression Model.