SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
This Algorithm Theoretical Basis (ATB) document describes the algorithms used to retrieve the cloud classification parameters of the MISR Level 2 Top-of-Atmosphere (TOA)/Cloud Product.
We trained two Random Forest (RF) machine-learning models for cloud mask and cloud thermodynamic phase detection using spectral observations from VIIRS on Suomi NPP (SNPP). Observations from CALIOP were carefully selected to provide reference labels. The two RF models were trained for all-day and daytime-only conditions using a 4-year collocated VIIRS/CALIOP dataset from 2013 to 2016. Due to the orbit difference, the collocated CALIOP and SNPP VIIRS training samples cover a broad viewing zenith angle range, which is a great benefit to overall model performance. The all-day model uses 3 VIIRS infrared (IR) bands (8.6,11, and 12 μm) and the daytime model uses 5 Near-IR (NIR) and Shortwave-IR (SWIR) bands (0.86, 1.24, 1.38, 1.64 and 2.25 μm) together with the 3 IR bands to detect clear, liquid water, and ice cloud pixels. Up to 7 surface types, namely, ocean/water, forest, cropland, grassland, snow/ice, barren/desert, and shrubland, were considered separately to enhance performance for both models. Detection of cloudy pixels and thermodynamic phase with the two RF models were compared against collocated CALIOP products from 2017. It is shown that, with a conservative screening process that excludes the most challenging cloudy pixels for passive remote sensing, the two RF models have high accuracy rates in comparison with the CALIOP reference for both cloud detection and thermodynamic phase. Other existing SNPP VIIRS and Aqua MODIS cloud mask and phase products are also evaluated, with results showing that the two RF models and the MODIS MYD06 optical property phase product are the top 3 algorithms with respect to lidar observations during the daytime. During the nighttime, the RF all-day model works best for both cloud detection and phase, in particular for pixels over snow/ice surfaces. The present RF models can be extended to other similar passive instruments if training samples can be collected from CALIOP or other lidars. However, the quality of reference labels and potential sampling issues that may impact model performance would need further attention.
The detection of mUltiple cloud layers using satellite observations is important for retrieval algorithms as well as climate applications. In this paper, we describe a relatively simple algorithm to detect multiple cloud layers and distinguish them from vertically-extended clouds. The algorithm can be applied to coincident passive sensors that derive both cloud-top pressure from the thermal infrared observations and an estimate of solar photon pathlength from UV, visible, or near-IR measurements. Here, we use data from the A-train afternoon constellation of satellites: cloud-top pressure, cloud optical thickness, the multi-layer flag from the Aqua MODerate-resolution Imaging Spectroradiometer (MODIS) and the optical centroid cloud pressure from the Aura Ozone Monitoring Instrument (OMI). For the first time, we use data from the CloudSat radar to evaluate the results of a multi-layer cloud detection scheme. The cloud classification algorithms applied with different passive sensor configurations compare well with each other as well as with data from CloudSat. We compute monthly mean fractions of pixels containing multi-layer and vertically-extended clouds for January and July 2007 at the OMI spatial resolution (l2kmx24km at nadir) and at the 5kmx5km MODIS resolution used for infrared cloud retrievals. There are seasonal variations in the spatial distribution of the different cloud types. The fraction of cloudy pixels containing distinct multi-layer cloud is a strong function of the pixel size. Globally averaged, these fractions are approximately 20% and 10% for OMI and MODIS, respectively. These fractions may be significantly higher or lower depending upon location. There is a much smaller resolution dependence for fractions of pixels containing vertically-extended clouds (approx.20% for OMI and slightly less for MODIS globally), suggesting larger spatial scales for these clouds. We also find higher fractions of vertically-extended clouds over land as compared with ocean, particularly in the tropics and summer hemisphere.
The structural characteristics of stratocumulus cloud fields off the coast of southern California are investigated using Landsat Multispectral Scanner imagery. Twelve scenes in this area are examined along with three other stratocumulus scenes near San Francisco, over central Oregon, and in the Gulf of Mexico. Results from this initial study of stratocumulus clouds indicate that cloud-background threshold selection techniques based upon edge detection gradient assumptions are not appropriate for cloud segmentation and classification algorithms, cloud size distributions obey a power law, and cell horizontal aspect ratio increases with cell diameter. It was also found that stratocumulus clouds are bifractal in nature with fractal dimension of d of about 1.2 for cells with diameter D smaller than 0.5 km and d of about 1.5 for cells with D greater than 0.5 km; stratocumulus cloud fields appear to be homogeneous over regions of about 100 km x 100 km, a much smaller region than the 2.5-deg x 2.5-deg boxes to be used in the ISCCP regional averaging algorithms; and that structural properties of stratocumulus clouds observed off the coast of southern California are similar to those observed for stratocumulus clouds at three other locations.
The Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES) was an aircraft-based campaign (15 January–26 February 2018) that deployed in situ probes and remote sensors to investigate low-level clouds over the Southern Ocean (SO). A novel methodology was developed to identify cloud boundaries and classify cloud phases in single-layer, low-level marine boundary layer (MBL) clouds below 3 km using the HIAPER Cloud Radar (HCR) and in situ measurements. The cloud base and top heights derived from HCR reflectivity, Doppler velocity, and spectrum width measurements agreed well with corresponding lidar-based and in situ estimates of cloud boundaries, with mean differences below 100 m. A liquid water content–reflectivity (LWC-Z) relationship, LWC = 0.70Z0.29, was derived to retrieve the LWC and liquid water path (LWP) from HCR profiles. The cloud phase was classified using HCR measurements, temperature, and LWP, yielding 40.6% liquid, 18.3% mixed-phase, and 5.1% ice samples, along with drizzle (29.1%), rain (3.2%), and snow (3.7%) for drizzling cloud cases. The classification algorithm demonstrates good consistency with established methods. This study provides a framework for the boundary and phase detection of MBL clouds, offering insights into SO cloud microphysics and supporting future efforts in satellite retrievals and climate model evaluation.
The new 3-year CloudSat and CALIPSO Science Team project described in this poster will use a unique combination of data from the Cloud-Aerosol LIdar with Orthogonal Polarization (CALIOP) instrument on CALIPSO and the Microwave Limb Sounder (MLS) on Aura, in conjunction with supporting meteorological information and detailed modeling studies, to advance our understanding of polar stratospheric cloud (PSC) processes and their role in ozone depletion. We will develop a third-generation (Gen3) PSC detection and composition algorithm that incorporates a new, more robust two-dimensional, multi-channel CALIOP feature detection scheme (2D-McDA). We will also devise and implement an improved two-dimensional PSC composition classification scheme that utilizes multiple parameters (e.g., CALIOP 532-nm parallel and perpendicular scattering ratios, CALIOP 1064-nm total scattering ratio, MLS HNO3 and H2O, ambient temperature, and temperature histories) in a Bayesian approach to determine the most likely PSC composition and help constrain solid PSC particle number density and size/shape. The combined CALIOP/MLS analyses will allow us to study in detail the full life cycle of PSCs and their resulting impact on gas-phase HNO3 and H2O, which should lead to improved parameterizations of PSC microphysics in global CCMs where detailed particle information is not available. The Gen3 CALIOP PSC algorithm will be a natural stepping-stone toward the analysis of data collected during future spaceborne lidar missions, such as NASA’s Atmosphere Observation System (AtmOS) mission currently scheduled for launch late in this decade. We will also investigate possible trends in PSC occurrence and composition over the entire CALIOP data record and through further comparisons with the Stratospheric Aerosol Measurement (SAM) II solar occultation PSC record from 1979-1989. Finally, we will validate the mountain-wave parameterization and PSC schemes used in the UM-UKCA (Unified Model coupled to the United Kingdom Chemistry and Aerosol module) chemistry-climate model through detailed comparisons with earlier CALIOP PSC data products and those developed under this proposal.
The new 3-year CloudSat and CALIPSO Science Team project described in this poster will use a unique combination of data from the Cloud-Aerosol LIdar with Orthogonal Polarization (CALIOP) instrument on CALIPSO and the Microwave Limb Sounder (MLS) on Aura, in conjunction with supporting meteorological information and detailed modeling studies, to advance our understanding of polar stratospheric cloud (PSC) processes and their role in ozone depletion. We will develop a third-generation (Gen3) PSC detection and composition algorithm that incorporates a new, more robust two-dimensional, multi-channel CALIOP feature detection scheme (2D-McDA). We will also devise and implement an improved two-dimensional PSC composition classification scheme that utilizes multiple parameters (e.g., CALIOP 532-nm parallel and perpendicular scattering ratios, CALIOP 1064-nm total scattering ratio, MLS HNO3 and H2O, ambient temperature, and temperature histories) in a Bayesian approach to determine the most likely PSC composition and help constrain solid PSC particle number density and size/shape. The combined CALIOP/MLS analyses will allow us to study in detail the full life cycle of PSCs and their resulting impact on gas-phase HNO3 and H2O, which should lead to improved parameterizations of PSC microphysics in global chemistry-climate models (CCMs) where detailed particle information is not available. The Gen3 CALIOP PSC algorithm will be a natural stepping-stone toward the analysis of data collected during future spaceborne lidar missions, such as NASA’s Atmosphere Observation System (AOS) mission currently scheduled for launch late in this decade. We will also investigate possible trends in PSC occurrence and composition over the entire CALIOP data record and through further comparisons with the Stratospheric Aerosol Measurement (SAM) II solar occultation PSC record from 1979-1989. Finally, we will validate the mountain-wave parameterization and PSC schemes used in the UM-UKCA (Unified Model coupled to the United Kingdom Chemistry and Aerosol module) CCM through detailed comparisons with earlier CALIOP PSC data products and those developed under this proposal.
To prepare for implementation of a new aerosol retrieval specifically designed for dust aerosol over ocean in the operational Dark-Target (DT) algorithms for the Moderate-resolution Imaging Spectrometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) satellite sensors, we focus on the challenge of detecting dust. We first survey the literature on existing dust detection algorithms and then develop an innovative algorithm that combines near-UV (deep blue), visible, and thermal infrared (TIR) wavelength spectral tests. The new detection algorithm is applied to Terra and Aqua MODIS granules and compared with other dust detection possibilities from existing MODIS products. Quantitative evaluation of the new dust detection algorithm is conducted using both a collocated AERONET-MODIS data set and collocated Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO)-MODIS data set. From comparison with both AERONET and CALIOP measurements, we estimate the new dust detection algorithm detects about 30% of weakly dusty pixels and more than 80% of heavily dusty pixels, with false detections in the range of 1–2%. The very low false detection rate is particularly noteworthy in comparison with existing literature. Compared with the dust flag currently available as part of the MODIS cloud mask product (MOD35/MYD35), and dust classification based on commonly used thresholds with aerosol optical depth (AOD) and Angstrom exponent (AE), the new dust detection algorithm finds more dusty pixels and fewer false detections.
In this paper we describe the algorithm hat will be used during the upcoming Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) mission for discriminating between clouds and aerosols detected in two wavelength backscatter lidar profiles. We first analyze single-test and multiple-test classification approaches based on one-dimensional and multiple-dimensional probability density functions (PDFs) in the context of a two-class feature identification scheme. From these studies we derive an operational algorithm based on a set of 3-dimensional probability distribution functions characteristic of clouds and aerosols. A dataset acquired by the Cloud Physics Lidar (CPL) is used to test the algorithm. Comparisons are conducted between the CALIPSO algorithm results and the CPL data product. The results obtained show generally good agreement between the two methods. However, of a total of 228,264 layers analyzed, approximately 5.7% are classified as different types by the CALIPSO and CPL algorithm. This disparity is shown to be due largely to the misclassification of clouds as aerosols by the CPL algorithm. The use of 3-dimensional PDFs in the CALIPSO algorithm is found to significantly reduce this type of error. Dust presents a special case. Because the intrinsic scattering properties of dust layers can be very similar to those of clouds, additional algorithm testing was performed using an optically dense layer of Saharan dust measured during the Lidar In-space Technology Experiment (LITE). In general, the method is shown to distinguish reliably between dust layers and clouds. The relatively few erroneous classifications occurred most often in the LITE data, in those regions of the Saharan dust layer where the optical thickness was the highest.
In recent years the availability of automated change detection algorithms in Google Earth Engine has permitted cloud-based processing of large time series of satellite imagery. Models such as the Continuous Change Detection and Classification (CCDC), CCDC-Spectral Mixture Analysis (CCDC-SMA), and Landsat-based Detection of Trends in Disturbance and Recovery (LandTrendr) allow users to exploit decades of Earth Observations (EO) , leveraging the Landsat archive and data from other sensors to detect disturbance in forest ecosystems. Despite the wide adoption of these methods, robust documentation and growing community of users, little research has explored their use in mangrove environments. Mangroves are dynamic environments subject to changes due to not only the natural migration of mudflats but also coastal erosion, urban expansion, aquaculture practices, etc. This work aims to identify best practices for the application of these models to identify and monitor changes in the Belizean mangroves, which experienced an estimated 5.4% decrease in extent between 1980 and 2017 (Cherrington et al. 2020). Partnering directly with the Belizean Forest Department, our team will develop a replicable, efficient methodology to annually update the country’s mangrove extent employing EO-based change detection. This collaboration supports Belize’s marine conservation and climate resilience efforts, such as its Blue Bond agreement with the Nature Conservancy, building on national processes to monitor these critical ecosystems.
The cloud cover in a set of summertime and wintertime AVHRR data from the Arctic and Antarctic regions was analyzed using a pattern recognition algorithm. The data were collected by the NOAA-7 satellite on 6 to 13 Jan. and 1 to 7 Jul. 1984 between 60 deg and 90 deg north and south latitude in 5 spectral channels, at the Global Area Coverage (GAC) resolution of approximately 4 km. This data embodied a Polar Cloud Pilot Data Set which was analyzed by a number of research groups as part of a polar cloud algorithm intercomparison study. This study was intended to determine whether the additional information contained in the AVHRR channels (beyond the standard visible and infrared bands on geostationary satellites) could be effectively utilized in cloud algorithms to resolve some of the cloud detection problems caused by low visible and thermal contrasts in the polar regions. The analysis described makes use of a pattern recognition algorithm which estimates the surface and cloud classification, cloud fraction, and surface and cloudy visible (channel 1) albedo and infrared (channel 4) brightness temperatures on a 2.5 x 2.5 deg latitude-longitude grid. In each grid box several spectral and textural features were computed from the calibrated pixel values in the multispectral imagery, then used to classify the region into one of eighteen surface and/or cloud types using the maximum likelihood decision rule. A slightly different version of the algorithm was used for each season and hemisphere because of differences in categories and because of the lack of visible imagery during winter. The classification of the scene is used to specify the optimal AVHRR channel for separating clear and cloudy pixels using a hybrid histogram-spatial coherence method. This method estimates values for cloud fraction, clear and cloudy albedos and brightness temperatures in each grid box. The choice of a class-dependent AVHRR channel allows for better separation of clear and cloudy pixels than does a global choice of a visible and/or infrared threshold. The classification also prevents erroneous estimates of large fractional cloudiness in areas of cloudfree snow and sea ice. The hybrid histogram-spatial coherence technique and the advantages of first classifying a scene in the polar regions are detailed. The complete Polar Cloud Pilot Data Set was analyzed and the results are presented and discussed.
Accurate, automated cloud and cloud shadow detection is a key component of the processing needed to prepare optical satellite imagery for scientific analysis. Many existing cloud detection algorithms rely on temperature information to identify clouds, making detection difficult for imagers that lack a thermal band, like Sentinel-2. To get maximum benefit from Sentinel-2 products it is critical to understand which algorithms best identify clouds and their shadows in images. We examined the relative performance of five different cloud-masking algorithms (Sen2Cor, MAJA, LaSRC, Fmask and Tmask) in 6 Sentinel-2 scenes (28 total images) distributed across the Eastern Hemisphere. Expanding on these comparisons, we tested ensemble approaches to improve results. We tested three ensemble approaches to cloud and shadow classification based on the outputs of the five initial algorithms using the cloud masks in: (1) a majority prediction model; (2) a random forests model; and (3) a conditional logic model. Accuracy assessments show a trade-off between omission and commission errors in cloud detection for individual algorithms across all sites, and some algorithms are better at detecting either clouds or cloud shadows. No single algorithm outperforms the others for both clouds and shadows. Aggregating the results from multiple algorithms produces fewer undetected clouds and higher overall accuracy than any single algorithm, with as high as 2.7% improvement over the top-performing algorithm, suggesting an ensemble approach may be the most useful for processing of Sentinel-2 data.
The accurate classification of aerosol types injected into the stratosphere is important to properly characterize their chemical and radiative impacts within the Earth climate system. The updated stratospheric aerosol subtyping algorithm used in the version 4.5 (V4.5) release of the Cloud Aerosol Lidar with Orthogonal Polarization (CALIOP) level 2 data products now delivers more comprehensive and accurate classifications than its predecessor. The original algorithm identified four aerosol subtypes for layers detected above the tropopause: volcanic ash, smoke, sulfate/other, and polar stratospheric aerosol (PSA). In the revised algorithm, sulfates are separately identified as a distinct, homogeneous subtype, and the diffuse, weakly scattering layers previously assigned to the sulfate/other class are recategorized as a fifth “unclassified” subtype. By making two structural changes to the algorithm and revising two thresholds, the V4.5 algorithm improves the ability to discriminate between volcanic ash and smoke from pyrocumulonimbus injections, improves the fidelity of the sulfate subtype, and more accurately reflects the uncertainties inherent in the classification process. The 532 nm lidar ratio for volcanic ash was also revised to a value more consistent with the current state of knowledge. This paper briefly reviews the previous version of the algorithm (V4.1 and V4.2) then fully details the rationale and impact of the V4.5 changes on subtype classification frequency for specific events where the dominant aerosol type is known based on the literature. Classification accuracy is best for volcanic ash due to its characteristically high depolarization ratio. Smoke layers in the stratosphere are also classified with reasonable accuracy, though during the daytime a substantial fraction are misclassified as ash. It is also possible for mixtures of ash and sulfate to be misclassified as smoke. The V4.5 sulfate subtype accuracy is less than that for ash or smoke, with sulfates being misclassified as smoke about one-third of the time. However, because exceptionally tenuous layers are now assigned to the unclassified subtype and the revised algorithm levies more stringent criteria for identifying an aerosol as sulfate, it is more likely that layers labeled as this subtype are in fact sulfate compared to those assigned the sulfate/other classification in the previous data release.
The retrieval of optimal aerosol datasets by the synergistic use of hyperspectral ultraviolet(UV)–visible and broadband meteorological imager (MI) techniques was investigated. The Aura Ozone Monitoring Instrument (OMI) Level 1B (L1B) was used as a proxy for hyperspectral UV–visible instrument data to which the Geostationary Environment Monitoring Spectrometer (GEMS) aerosol algorithm was applied. Moderate-Resolution Imaging Spectroradiometer (MODIS) L1B and dark target aerosol Level 2 (L2) data were used with a broadband MI to take advantage of the consistent time gap between the MODIS and the OMI. First, the use of cloud mask information from the MI infrared (IR) channel was tested for synergy. High-spatial-resolution and IR channels of the MI helped mask cirrus and sub-pixel cloud contamination of GEMS aerosol, as clearly seen in aerosol optical depth (AOD) validation with Aerosol Robotic Network (AERONET) data. Second, dust aerosols were distinguished in the GEMS aerosol-type classification algorithm by calculating the total dust confidence index (TDCI) from MODIS L1B IR channels. Statistical analysis indicates that the Probability of Correct Detection (POCD) between the forward and inversion aerosol dust models (DS) was increased from 72% to 94% by use of the TDCI for GEMS aerosol-type classification, and updated aerosol types were then applied to the GEMS algorithm. Use of the TDCI for DS type classification in the GEMS retrieval procedure gave improved single-scattering albedo (SSA) values for absorbing fine pollution particles (BC) and DS aerosols. Aerosol layer height (ALH) retrieved from GEMS was compared with Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) data, which provides high-resolution vertical aerosol profile information. The CALIOP ALH was calculated from total attenuated backscatter data at 1064 nm, which is identical to the definition of GEMS ALH. Application of the TDCI value reduced the median bias of GEMS ALH data slightly. The GEMS ALH bias approximates zero, especially for GEMS AOD values of>~0.4 and GEMS SSA values of<~0.95.Finally, the AOD products from the GEMS algorithm and MI were used in aerosol merging with the maximum-likelihood estimation method, based on a weighting factor derived from the standard deviation of the original AOD products. With the advantage of the UV–visible channel in retrieving aerosol properties over bright surfaces, the combined AOD products demonstrated better spatial data availability than the original AOD products, with comparable accuracy. Furthermore, pixel-level error analysis of GEMS AOD data indicates improvement through MI synergy.
NASA's Ice, Cloud and Land Elevation Satellite-II (ICESat-2) mission is a decadal survey mission (2016 launch). The mission objectives are to measure land ice elevation, sea ice freeboard, and changes in these variables, as well as to collect measurements over vegetation to facilitate canopy height determination. Two innovative components will characterize the ICESat-2 lidar: 1) collection of elevation data by a multibeam system and 2) application of micropulse lidar (photon-counting) technology. A photon-counting altimeter yields clouds of discrete points, resulting from returns of individual photons, and hence new data analysis techniques are required for elevation determination and association of the returned points to reflectors of interest. The objective of this paper is to derive an algorithm that allows detection of ground under dense canopy and identification of ground and canopy levels in simulated ICESat-2 data, based on airborne observations with a Sigma Space micropulse lidar. The mathematical algorithm uses spatial statistical and discrete mathematical concepts, including radial basis functions, density measures, geometrical anisotropy, eigenvectors, and geostatistical classification parameters and hyperparameters. Validation shows that ground and canopy elevation, and hence canopy height, can be expected to be observable with high accuracy by ICESat-2 for all expected beam energies considered for instrument design (93.01%-99.57% correctly selected points for a beam with expected return of 0.93 mean signals per shot (msp), and 72.85%-98.68% for 0.48 msp). The algorithm derived here is generally applicable for elevation determination from photoncounting lidar altimeter data collected over forested areas, land ice, sea ice, and land surfaces, as well as for cloud detection.
The Feature Identification and Location Experiment (FILE) that is being designed for the detection and classification of four primary earth features (water, vegetation, bare land, and the clouds-snow-ice class) is described. Consideration is given to the FILE classification technology concept and the FILE instrument, which will use two solid-state CCD cameras operating at 0.65 and 0.85-micron center frequency wavelengths, with the camera outputs being functions of the earth surface material radiance. The classification is based on camera output radiance ratio values. The preliminary analysis of the data collected on the STS 41-G mission is discussed. The results demonstrated the suitability of using the two-channel-ratio detection technology and a simple (y = mx) algorithm to autonomously classify the four earth surface features. The technology is especially attractive as a cloud sensor, where, in advance of or during a mission, a threshold value for cloud cover percentage can be programmed and/or adaptively modified for use in the control of other remote sensors.
The detection of inland water bodies from Synthetic Aperture Radar (SAR) data provides a great advantage over water detection with optical data, since SAR imaging is not impeded by cloud cover. Traditional methods of detecting water from SAR data involves using thresholding methods that can be labor intensive and imprecise. This paper describes Water Across Synthetic Aperture Radar Data (WASARD): a method of water detection from SAR data which automates and simplifies the thresholding process using machine learning on training data created from Geoscience Australia’s WOFS algorithm. Of the machine learning models tested, the Linear Support Vector Machine was determined to be optimal, with the option of training using solely the VH polarization or a combination of the VH and VV polarizations. WASARD was able to identify water in the target area with a correlation of 97% with WOFS. Sentinel-1, Open Data Cube, Earth Observations, Machine Learning, Water Detection 1. INTRODUCTION Water classification is an important function of Earth imaging satellites, as accurate remote classification of land and water can assist in land use analysis, flood prediction, climate change research, as well as a variety of agricultural applications [2]. The ability to identify bodies of water remotely via satellite is immensely cheaper than contracting surveys of the areas in question, meaning that an application that can accurately use satellite data towards this function can make valuable information available to nations which would not be able to afford it otherwise. Highly reliable applications for the remote detection of water currently exist for use with optical satellite data such as that provided by LANDSAT. One such application, Geoscience Australia’s Water Observations from Space (WOFS) has already been ported for use with the Open Data Cube [6]. However, water detection using optical data from Landsat is constrained by its relatively long revisit cycle of 16 days [5], and water detection using any optical data is constrained in that it lacks the ability to make accurate classifications through cloud cover [2]. The alternative solution which solves these problems is water detection using SAR data, which images the Earth using cloud-penetrating microwaves. Because of its advantages over optical data, much research has been done into water detection using SAR data. Traditionally, this has been done using the thresholding method, which involves picking a polarization band and labeling all pixels for which this band’s value is below a certain threshold as containing water. The thresholding method works since water tends to return a much lower backscatter value to the satellite than land [1]. However, this method can be flawed since estimating the proper threshold is often imprecise, complicated, and labor intensive for the end user. Thresholding also tends to use data from only one SAR polarization, when a combination of polarizations can provide insight into whether water is present. [2] In order to alleviate these problems, this paper presents an application for the Open Data Cube to detect water from SAR data using support vector machine (SVM) classification. 2. PLATFORM WASARD is an application for the Open Data Cube, a mechanism which provides a simple yet efficient means of ingesting, storing, and retrieving remote sensing data. Data can be ingested and made analysis ready according to whatever specifications the researcher chooses, and easily resampled to artificially alter a scene’s resolution. Currently WASARD supports water detection on scenes from ESA’s Sentinel-1 and JAXA’s ALOS. When testing WASARD, Sentinel-1 was most commonly used due to its relatively high spatial resolution and its rapid 6 day revisit cycle [5]. With minor alterations to the application's code, however, it could support data from other satellites. 3. METHODOLOGY Using supervised classification, WASARD compares SAR data to a dataset pre-classified by WOFS in order to train an SVM classifier. This classifier is then used to detect water in other SAR scenes outside the training set. Accuracy was measured according to the following metrics: Precision: a measure of what percentage of the points WASARD labels as water are truly water Recall: a measure of what percentage of the total water cover WASARD was able to identify. F1 Score: a harmonic average of the precision and recall scores Both precision and recall are calculated at the end of the training phase, when the trained classifier is compared to a testing dataset. Because the WOFS algorithm’s classifications are used as the truth values when training a WASARD classifier, when precision and recall are mentioned in this paper, they are always with respect to the values produced by WOFS on a similar scene of Landsat data, which themselves have a classification accuracy of 97% [6]. Visual representations of water identified by WASARD in this paper were produced using the function wasard_plot(), which is included in WASARD. 3.1 Algorithm Selection The machine learning model used by WASARD is the Linear Support Vector Machine (SVM). This model uses a supervised learning algorithm to develop a classifier, meaning it creates a vector which can be multiplied by the vector formed by the relevant data bands to determine whether a pixel in a SAR scene contains water. This classifier is trained by comparing data points from selected bands in a SAR scene to their respective labels, which in this case are “water” or “not water” as given by the WOFS algorithm. The SVM was selected over the Random Forest model, which outperformed the SVM in training speed, but had a greater classification time and lower accuracy, and the Multilayer Perceptron Artificial Neural Network, which had a slightly higher average accuracy than the SVM, but much greater training and classification times. Figure 1: Visual representation of the SVM Classifier. Each white point represents a pixel in a SAR scene. In Figure 1, the diagonal line separating pixels determined to be water from those determined not to be water represents the actual classification vector produced by the SVM. It is worth noting that once the model has been trained, classification of pixels is done in a similar manner as in the thresholding method. This is especially true if only one band was used to train the model. 3.1 Feature Selection Sentinel-1 collects data from two bands: the Vertical/Vertical polarization (VV) and the Vertical/Horizontal polarization (VH). When 100 SVM classifiers were created for each polarization individually, and for the combination of the two, the following results were achieved: Figure 2: Accuracy of classifiers trained using different polarization bands. Precision and Recall were measured with respect to the values produced by WOFS. Figure 2 demonstrates that using both the VV and VH bands trades slightly lower recall for significantly greater precision when compared with the VH band alone, and that using the VV band alone is inferior in both metrics. WASARD therefore defaults to using both the VV and VH bands, and includes the option to use solely the VH band. The VV polarization’s lower precision compared to the VH polarization is in contrast to results from previous research and may merit further analysis [4]. 3.2 Training a Classifier The steps in training a classifier with WASARD are 1. Selecting two scenes (one SAR, one optical) with the same spatial extents, and acquired close to each other in time, with a preference that the scenes are taken on the same day. 2. Using the WOFS algorithm to produce an array of the detected water in the scene of optical data, to be used as the labels during supervised learning 3. Data points from the selected bands from the SAR acquisition are bundled together into an array with the corresponding labels gathered from WOFS. A random sample with an equal number of points labeled “Water” and “Not Water” is selected to be partitioned into a training and a testing dataset 4. Using Scikit-Learn’s LinearSVC object, the training dataset is used to produce a classifier, which is then tested against the testing dataset to determine its precision and recall The result is a wasard_classifier object, which has the following attributes: 1. f1, recall, and precision: 3 metrics used to determine the classifier’s accuracy 2. Coefficient: Vector which the SVM uses to make its predictions. The classifier detects water when the dot product of the coefficient and the vector formed by the SAR bands is positive 3. Save(): allows a user to save a classifier to the disk in order to use it without retraining 4. wasard_classify(): Classifies an entire xarray of SAR data using the SVM classifier All of the above steps are performed automatically when the user creates a wasard_classifier object. 3.3 Classifying a Dataset Once the classifier has been created, it can be used to detect water in an xarray of SAR data using wasard_classify(). By taking the dot product of the classifier’s coefficients and the vector formed by the selected bands of SAR data, an array of predictions is constructed. A classifier can effectively be used on the same spatial extents as the ones where it was trained, or on any area with a similar landscape. While
Multi-Angle Implementation of Atmospheric Correction (MAIAC) is a new algorithm developed for MODIS. MAIAC uses a time series analysis and processing of groups of pixels to perform simultaneous retrievals of aerosol properties and surface bidirectional reflectance without typical assumptions about the surface. It is a generic algorithm which works over both dark and bright land surfaces, including deserts. MAIAC has an internal Cloud Mask, a dynamic land-water-snow classification and a surface change mask which allows it to flexibly choose processing path over different surfaces. A distinct feature of MAIAC is a high 1 km resolution of aerosol retrievals which is required in different applications including the air quality analysis. The novel features of MAIAC include the high quality cloud mask, discrimination of aerosol type, including biomass burning smoke and dust, and detection of surface change - all required for high quality aerosol retrievals. An overview of the algorithm, results of AERONET validation, and examples of comparison with MODIS Collection 5 aerosol product and Deep Blue algorithm for different parts of the world, will be presented.