Simple Atmospheric Correction Research Articles

Using Landsat to extend the historical record of lacustrine phytoplankton blooms: A Lake Erie case study

Long-term records of phytoplankton blooms in freshwater lakes are necessary both for understanding basin-scale changes to watersheds and for providing a key constraint for assessing processes driving blooms. However, due to the inherent constraints of in situ sampling and the short time period covered by modern space borne sensors, few long-term records exist. Historical data from sensors such as Landsat offer strong potential for creating new records of past blooms. Here, we use a novel evaluation procedure based on multiple metrics to assess algorithm suitability and robustness for generating long-term bloom records using Landsat 5 imagery. Evaluation metrics are based on bloom presence, spatial distribution, magnitude and timing, using both in situ Microcystis biovolume and remotely-sensed Cyanobacterial Index (CI) data from MERIS for 2002–2011. Applying this procedure for a test case focusing on Lake Erie's western basin, an algorithm based on a near infrared threshold with simple atmospheric correction through subtraction of the shortwave infrared band, combined with an additional “greenness” filter based on a hue threshold, performs best. Implementing this algorithm for 1984–2001 reveals the long-term trends in peak bloom magnitude prior to the start of the MERIS and MODIS record (2002–2015), and more than doubles the period of record that can be used to understand bloom occurrence and growth for this system. More broadly, we demonstrate that Landsat observations can be used to identify macro-scale features of blooms. For Lake Erie, the performance of the final Landsat algorithm is comparable to that of the MERIS CI algorithm, despite Landsat's broad spectral bands and long revisit time.

AHS and CASI Processing for the REFLEX Remote Sensing Campaign: Methods and Results

The airborne spectroradiometers AHS and CASI were used as a source of hyperspectral and thermal remote sensing data during the REFLEX campaign. Data geolocation and a first simple atmospheric correction was performed by INTA in near-real time with a specific on-site setup and distributed to all campaign participants. In this paper we present briefly the AHS and CASI REFLEX flight campaign followed by a detailed description of the methodology used for image processing and finally the results obtained in terms of image quality. As a conclusion, near-real time processing for AHS and CASI level 1 geolocated products was successful as most of CASI level 2 results but further work is needed for achieving accurate AHS level 2 products.

Temporal generalization of sub-pixel vegetation mapping with multiple machine learning and atmospheric correction algorithms

Temporal generalization allows a trained classification algorithm to be applied to multiple images across time to derive reliable classification map products. It is a challenging remote-sensing research topic since the results are dependent on the selection of atmospheric correction methods, classification algorithms, validation processes, and their varying combinations. This study examined the temporal generalization of sub-pixel vegetation mapping using multiple Landsat images (1990, 1996, 2004, and 2010). All Landsat images were processed with two atmospheric correction methods: simple dark object subtraction (DOS) and the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) algorithm. For the sub-pixel vegetation mapping of the 2004 Landsat image, we used high-resolution OrbView-3 images as a training/validation data set and compared three machine learning algorithms (neural networks, random forests, and classification and regression trees) for their classification performance. The trained classifiers were then applied to other Landsat images (1990, 1996, and 2010) to derive sub-pixel vegetation map products. For the 2004 Landsat image classification, cross-validation shows similar classification results for neural networks (root mean square error (RMSE) = 0.099) and random forests (RMSE = 0.100) algorithms, and both are better than classification and regression trees (RMSE = 0.123). Pseudo-invariant pixels between 2004 and 2010 were used as validation points to evaluate the temporal generalizability of classification algorithms. Simple DOS and LEDAPS atmospheric correction resulted in similar accuracy statistics. The neural-network-based classifier performed best in generating reliable sub-pixel vegetation map products across time.

A Simple Atmospheric Correction Algorithm for MODIS in Shallow Turbid Waters: A Case Study in Taihu Lake

A simple atmospheric correction algorithm (SACA) based on a “known” empirical spectral relationship at two bands of MODIS sensor has been developed as a modification to improve the MODIS ocean color products in shallow turbid waters. The analysis results for spectral characteristics suggest that 531, 551, 667, and 678 nm are the optimal bands for constructing the empirical spectral relationship of the SACA algorithm. In order to work efficiently with high adjacency effects and water bottom reflectance in shallow inland lakes, the image statistics method is suggested for the SACA algorithm to determine the aerosol scattering contribution. Using the in situ measurements taken in Taihu Lake on October 27 and 28, 2003, MODIS water-leaving reflectance derived from the atmospheric correction were evaluated using the SWIR and SACA methods, indicating that the SACA algorithm produces a superior performance at visible bands but provides a poor result at NIR bands to the SWIR algorithm. To solve this problem, a combined method is suggested: the SACA algorithm is operated at visible bands, while the SWIR algorithm is executed at NIR bands. Results indicate that the MODIS-derived water-leaving reflectance in shallow turbid water within 20% may be obtained using the combined method. These findings imply that the SACA algorithm is an acceptable atmospheric correction method for deriving the water-leaving reflectance from MODIS data in shallow turbid waters.

Estimation of discharges at river mouth with MODIS image

The transport of the sediment, carried in suspension by water, is central to hydrology and the ecological functioning of river floodplains and deltas. River discharge estimation is useful for demonstrating this information. In this study, we extracted MODIS reflectance values from a pixel near the river mouth after carrying out the simple atmospheric correction method, then applied single regression analysis to reflectance values and the in situ discharge of Naka River in Tokushima prefecture and Monobe River in Kochi prefecture, Japan. MODIS images and in situ data were taken from January through December, 2004. As a result, both in Naka River and Monobe River, robustly positive relationships between the discharges observed in situ and remotely sensed MODIS reflectance data in the region of river mouth were found throughout the year. In addition, we estimated monthly and annual average discharge from the MODIS reflectance with the regression formula. As a result, in situ average discharge was well estimated.

MERIS Retrieval of Water Quality Components in the Turbid Albemarle-Pamlico Sound Estuary, USA

Two remote-sensing optical algorithms for the retrieval of the water quality components (WQCs) in the Albemarle-Pamlico Estuarine System (APES) were developed and validated for chlorophyll a (Chl). Both algorithms were semi-empirical because they incorporated some elements of optical processes in the atmosphere, water, and air/water interface. One incorporated a very simple atmospheric correction and modified quasi-single-scattering approximation (QSSA) for estimating the spectral Gordon’s parameter, and the second estimated WQCs directly from the top of atmosphere satellite radiance without atmospheric corrections. A modified version of the Global Meteorological Database for Solar Energy and Applied Meteorology (METEONORM) was used to estimate directional atmospheric transmittances. The study incorporated in situ Chl data from the Ferry-Based Monitoring (FerryMon) program collected in the Neuse River Estuary (n = 633) and Pamlico Sound (n = 362), along with Medium Resolution Imaging Spectrometer (MERIS) satellite imagery collected (2006–2009) across the APES; providing quasi-coinciding samples for Chl algorithm development and validation. Results indicated a coefficient of determination (R2) of 0.70 and mean-normalized root-mean-squares errors (NRMSE) of 52% in the Neuse River Estuary and R2 = 0.44 (NRMSE = 75 %) in the Pamlico Sound—without atmospheric corrections. The simple atmospheric correction tested provided on performance improvements. Algorithm performance demonstrated the potential for supporting long-term operational WQCs satellite monitoring in the APES.

Water quality mapping using digital camera images

The objective of this study was to implement a low-cost airborne remote sensing system to provide an alternative solution for remote sensing given the difficulty in obtaining cloud-free satellite scenes of the equatorial region. An inexpensive sensor, a Kodak DC 290 digital camera, was used, onboard a light aircraft, Cessna 172Q. The feasibility of using camera images for remote sensing applications was tested for quantifying total suspended solids (TSS) from three estuaries located in the northern region of Peninsular Malaysia. The aircraft was flown at altitudes of 914.4 to 2438.4 m for digital image acquisition of the study areas. Water samples were collected simultaneously with the aircraft overpasses and their locations were determined by using a hand-held Global Positioning System (GPS). Oblique images were corrected for brightness variation. A simple relative atmospheric correction was performed on the multidate images. The captured colour images were then separated into three bands (red, green and blue) for multispectral analysis. The digital numbers (DNs) were extracted corresponding to the sea data locations for each band. A multiband regression algorithm was developed based on a reflectance model, which is a function of the inherent optical properties of water, and this in turn can be related to the concentration of the TSS. Data from different scenes were combined and then divided into two sets, one for calibration of the algorithm and the other for a validation analysis. The calibrated algorithm had a correlation coefficient (R) of 0.96 and a root mean square error (RMSE) of approximately 17 mg l−1. The validation analysis showed that the algorithm could estimate the TSS concentration within an RMSE of about 23 mg l−1 and an R value of 0.95. The calibrated algorithm was used to generate water quality maps for all images. The maps were geometrically corrected and colour coded for visual interpretation.

The use of selected pseudo-invariant targets for the application of atmospheric correction in multi-temporal studies using satellite remotely sensed imagery

Because atmospheric effects can have a significant impact on the data obtained from multi-spectral satellite remote sensing, it is frequently necessary to make corrections before any other image processing can be started. This paper describes a robust and relatively simple atmospheric correction method that uses pseudo-invariant targets (PITs) in conjunction with the empirical line method. The method is based on the selection of a number of suitable generic PITs, on the basis that they are large, distinctive in shape, and occur in many geographical areas. Whereas the multi-temporal normalization method corrects all images to a selected reference image, in this method images are simultaneously corrected using targets with a range of estimated surface reflectance values. The paper describes some applications of the method for a range of environmental studies involving water quality and air pollution monitoring, and mapping land-cover changes.

Ground Deformation Detection at Kuchinoerabujima Volcano by Continuous GPS with Simple Atmospheric Correction

Ground Deformation Detection at Kuchinoerabujima Volcano by Continuous GPS with Simple Atmospheric Correction

Error Analysis of Retrieved Aerosol Optical Depth due to Adjacency Effect for ROCSAT-2 RSI Bands

The adjacency effect (AE), caused by the scattering of reflected radiance from an inhomogeneous surface, is known to introduce a blurring effect on remotely sensed data, especially for high spatial resolution data. The retrieval of aerosol optical depth (AOD) by the dark target (DT) method usually neglects the AE. Hence, significant errors may be induced in aerosol retrieval. In this study, the errors of retrieved AOD due to AE by the DT method for very high spatial resolution data of 8 m in the multi-spectral bands of ROCSAT-2 RSI are investigated. A fast, accurate simple atmospheric correction model is applied. Target size is considered the same as the spatial resolution. The errors are mainly studied for both low and medium contrasts, in both clear and hazy skies for both continental and urban aerosol models. The results indicate that for the continental model, AE may cause significant retrieval errors for AOD over DTs for medium contrast in both skies in blue and red bands. Error becomes negligible if the inhomogeneous surface is for low contrast in clear sky. It is larger for the urban model where there is more absorption than for the continental model, due to the less sensitivity of top-of-atmosphere reflectance to AOD for DTs in inhomogeneous surfaces. This suggests that AE be considered for medium contrast in both skies, especially for urban-industrial regions for ROCSAT2 RSI.

Synoptic water clarity assessment in the Florida Keys using diffuse attenuation coefficient estimated from Landsat imagery

The diffuse attenuation coefficient, K (m− 1), is a measure of the effective attenuation of light in the water column. It characterizes water clarity and is used as a proxy for water quality. Mapping of shallow water benthic habitats using optical means, including daytime visible satellite imagery, requires knowledge of K to correct for water column effects such as light absorption and scattering. Traditionally, K is derived from imagery using a priori knowledge of bottom types at different depths and specific locations, and assuming that light attenuates exponentially with depth. This technique is applied to three Landsat 7 satellite images (February, May, and July 2000) from the Florida Keys Reef Tract between Key Largo and Key West. Interpolated depth data, initially from NOAA vector chart data, with uncertainties of ±0.5 m, were draped over 30 m spatial resolution Landsat satellite imagery. K was derived for 27 sites where bright sand could be observed at depths between 3 and 20 m. The blue and green bands (Landsat bands 1 and 2 at 450–520 nm and 520–600 nm, respectively) provided K values consistent with time and location. Average K values for bands 1 and 2, respectively, were 0.029 and 0.043 m− 1 (Lower Keys), 0.050 and 0.072 m− 1 (Middle Keys), and 0.063 and 0.082 m− 1 (Upper Keys). The red band (band 3, 630–690 nm) provided more ambiguous and erroneous results with several negative K values, attributed primarily to three factors use of a simple atmospheric correction, the high absorption and rapid attenuation of red light in water, and low radiometric sensitivity of the Landsat ETM + sensor. Despite the fact that these observations were a snapshot in time, trends were observed for the Upper, Middle, and Lower Keys, possibly due to the influence of more turbid Florida Bay waters.

Assessment of SeaWiFs atmospheric and marine products for the Northern Adriatic Sea

An evaluation of the accuracy of atmospheric and marine satellite-derived products is presented and discussed for the northern Adriatic Sea coastal region using match-ups of in situ and Sea-Viewing Wide-Field-of-View Sensor (SeaWiFS) data for the period September 1997-September 2001. The study, making use of a simple atmospheric correction scheme including a near-infrared (NIR) turbid-water correction, has shown mean relative percentage differences between in situ and satellite-derived aerosol optical thickness lower than 23% in the spectral range between 443 and 865 nm. By applying regional empirical bio-optical algorithms for chlorophyll a concentration (Chla), total suspended matter concentration (TSM), and diffuse attenuation coefficient at 490 nm (K/sub d/(490)), match-ups analysis has shown mean relative percentage differences of 40% for Chla, 28% for TSM, and 30% for K/sub d/(490). The analysis is supported by comparison of in situ and satellite-derived normalized water leaving radiances to highlight the importance of the NIR turbid-water correction and to discuss the intrinsic uncertainties due to the use of empirical algorithms.

Comparing time series of AVHRR LAC data with the IGBP 1 km and the Pathfinder 8 km datasets

Time series of remote sensing vegetation index data are increasingly being used for environmental monitoring and assessments. An important issue is therefore to ensure the integrity in time of the data. The present study compares and assesses the consistency and quality of three different sets of NDVI data from NOAA AVHRR: 1) full resolution, single-day Local Area Coverage (LAC) data acquired by local receiving stations, 2) 10 day composites of LAC data compiled by the IGBP (International Geosphere-Biosphere Program) and 3) 10 day composites of Pathfinder AVHRR Land (PAL) 8 km data. The comparison covers Senegal for the period from 1990 to 1992 in order to assess the appropriateness of using either PAL or IGBP data alone for long-term time series analysis, as well as using PAL and IGBP data as a supplement to the local LAC database. A high level of consistency was found between the three datasets despite differences in spatial and temporal resolution, composition and differences in the chosen atmospheric correction for each dataset. All data captured the same general trend in NDVI through the growing season and a linear relationship was established between the LAC and PAL data within the season, indicating that the simple atmospheric correction applied to the PAL data was very static. All three datasets showed comparable data dynamics. Clouds and haze seriously contaminated the IGBP LAC data during the peak of the rainy season, especially since the number of individual scenes used for each maximum value composite was limited to 1 to 3 scenes. It was concluded that the PAL 8km data can be used independently for time series analysis and can be used to fill gaps in time-series of LAC data where a scale of 8? 8 km is appropriate. Nonetheless the full resolution LAC data were superior to the PAL data owing to better spatial and temporal resolution, a superior geometric registration and more stringent quality checks.

A simple atmospheric correction model for ROCSAT-2 RSI data

A simple atmospheric correction model (SACM) for ROCSAT-2 RSI data is presented. Gaseous transmission and Rayleigh optical depth are simplified as analytic functions. Rayleigh and aerosol scattering are determined using lookup tables. Results indicate the SACM not only quite accurately reproduces top-of-atmosphere reflectance computed by the 6S model (Vermote et al. 1997), but also does so faster than 6S. The results of the inverse application of SACM also indicate that the error of surface reflectance can be greatly reduced even in hazy sky if the adjacency effect is corrected. Future studies and possible improvements are also highlighted.

High-resolution ground verification, cluster analysis and optical model of reef substrate coverage on Landsat TM imagery (Red Sea, Egypt)

A combination of high-resolution ground verification, cluster analysis using Landsat Thematic Mapper (TM) data, and optical modelling, was applied to Red Sea reef substrate. Ground verification, in an area of 3 by 20 pixels (90 by 600 m) with one metre scale resolution, identified the presence of 30 different bottom types that were later reduced to twelve dominant bottom types. A combination of bispectral plots and principal component analysis using spectral bands 1, 2 and 3 confirmed the presence of nine bottom types. The identified clusters were separated and used as a training set to classify substrate. Optical modelling, using literature radiance values and coverage of the original twelve dominant bottom types and a simple model for atmospheric and water column absorption, revealed a difference of up to 60 W m -2 between predicted substrate radiance and the satellite sensor values in the reef top area. Considering the simple atmospheric correction model, the lack of in situ radiance measurements and the uncertainties with respect to possible changes in bottom type distribution since the acquisition of the 14 year old image, the results show the potential use of satellite imagery for reef research in both biological and geological analysis through very precise and semi-quantitative ground verification, including in situ reflectance measurements.

Evaluation of simplified procedures for retrieval of land surface reflectance factors from satellite sensor output

In response to the need for a simple atmospheric correction method and the consequent verification of such a method, an experiment was conducted to acquire a data set suitable for testing atmospheric correction procedures under a variety of atmospheric conditions. Several procedures, including radiative transfer codes (RTCs) with simulated atmospheres, image-based procedures and dark-object subtraction (DOS), were evaluated by comparing surface reflectance factors derived from Landsat Thematic Mapper (TM) digital data with low-altitude, aircraft-based measurements for seven dates over a 1-year period. Acceptable results, approximately ± 0.02 reflectance (1 σ RMS), were achieved based on an RTC with appropriate simulated atmospheres. The DOS technique was the least accurate method and, in fact, produced greater error in estimations of near-IR reflectance than no correction at all. Two hybrid approaches, which combined the image-based nature of DOS with the precision of an RTC, provided sufficient accuracy and simplicity to warrant consideration for use on an operational basis. Though these results were probably site-specific (characterized by relatively low aerosol levels and low humidity), they illustrate the feasibility of simple atmospheric correction methods and the usefulness of a diverse data set for validation of such techniques.

A simple atmospheric correction method for the short wave satellite images

A simple and fast correction method was developed to partly eliminate the blurring effects of the atmosphere on satellite imagery of the Earth's surface taken in the solar spectrum. This method is fast even on a small, low capacity computer. We show on Landsat images, that our correction procedure improves the contrast of the images and the results of the cluster analysis.

Albedo of the U.S. Great Plains as Determined from NOAA-9 AVHRR Data

Abstract The seasonal variation of surface albedo is derived from NOAA-9 AVHRR observations of the US. Great Plains during the snow-free months of 1986 and 1987. Monthly albedo maps are constructed using a simple model-independent technique which includes recording of the cloud-free data by 9-day satellite repeat cycle, spatial interpolation of the data averaging on each day of the cycle, and averaging over the cycle to obtain a monthly mean estimate. The surface albedo is obtained by applying a narrowband-to-broadband conversion and a simple atmospheric correction to the clear-sky albodo of each mouth. The phenological changes of four target areas (two forest lands and two croplands) are analyzed. In some areas, crop development can double the albedo values within the growing season. Such changes are important in radiation budget calculations. On the whole, the results are in good agreement with tabulated values based on albedos reported in the literature for various surface types.

A Simple Method for Estimating Monthly Mean Albedo of Land Surfaces From AVHRR Data

It is suggested that the observed periodicity of cloud-free, visible and near-infrared data from the Advanced Very High Resolution Radiometer (AVHRR) onboard NOAA-9 for a particular target can be efficiently used for deriving the monthly mean clear-sky planetary albedo. The broadband albedo is approximated by a linear combination of visible and near-infrared albedos. The new method of 9-day compositing distributes the clear-sky observations over the sun-target-sensor geometry combinations and weights them by the occurrence of each combination during a particular month. It is shown that the monthly mean clear-sky planetary albedo can be estimated solely from the visible and near-infrared data in a model-independent manner. The surface albedo is then obtained by applying a simple atmospheric correction to the derived clear-sky planetary albedo. A comparison of the results of the present method and those obtained by using the “minimum albedo” method indicates that the latter may lead to an underestimation of monthly values because of the angular variation in clear-sky observed albedos. The derived surface albedos are compared with those reported in climatological studies based on ground observations.