Land cover monitoring using remotely sensed data requires robust classification methods which allow for the accurate mapping of complex land cover and land use categories. Random forest (RF) is a powerful machine learning classifier that is relatively unknown in land remote sensing and has not been evaluated thoroughly by the remote sensing community compared to more conventional pattern recognition techniques. Key advantages of RF include: their non-parametric nature; high classification accuracy; and capability to determine variable importance. However, the split rules for classification are unknown, therefore RF can be considered to be black box type classifier. RF provides an algorithm for estimating missing values; and flexibility to perform several types of data analysis, including regression, classification, survival analysis, and unsupervised learning.

In this paper, the performance of the RF classifier for land cover classification of a complex area is explored. Evaluation was based on several criteria: mapping accuracy, sensitivity to data set size and noise. Landsat-5 Thematic Mapper data captured in European spring and summer were used with auxiliary variables derived from a digital terrain model to classify 14 different land categories in the south of Spain. Results show that the RF algorithm yields accurate land cover classifications, with 92% overall accuracy and a Kappa index of 0.92. RF is robust to training data reduction and noise because significant differences in kappa values were only observed for data reduction and noise addition values greater than 50 and 20%, respectively. Additionally, variables that RF identified as most important for classifying land cover coincided with expectations. A McNemar test indicates an overall better performance of the random forest model over a single decision tree at the 0.00001 significance level.

An assessment of the effectiveness of a random forest classifier for land-cover classification

A multilayer photonic structure is described that strongly reflects incident sunlight while emitting heat selectively through an atmospheric transparency window to outer space; this leads to passive cooling under direct sunlight of 5 degrees Celsius below ambient air temperature, which has potential applications in air-conditioning and energy efficiency.

/pdf/passive-radiative-cooling-below-ambient-air-temperature-4ldemmkn87.pdf

Passive radiative cooling below ambient air temperature under direct sunlight

Imaging Spectroscopy and the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS)

The Moderate Resolution Imaging Spectroradiometer (MODIS) is one of five instruments aboard the Terra Earth Observing System (EOS) platform launched in December 1999. After achieving final orbit, MODIS began Earth observations in late February 2000 and has been acquiring data since that time. The instrument is also being flown on the Aqua spacecraft, launched in May 2002. A comprehensive set of remote sensing algorithms for cloud detection and the retrieval of cloud physical and optical properties have been developed by members of the MODIS atmosphere science team. The archived products from these algorithms have applications in climate change studies, climate modeling, numerical weather prediction, as well as fundamental atmospheric research. In addition to an extensive cloud mask, products include cloud-top properties (temperature, pressure, effective emissivity), cloud thermodynamic phase, cloud optical and microphysical parameters (optical thickness, effective particle radius, water path), as well as derived statistics. We will describe the various algorithms being used for the remote sensing of cloud properties from MODIS data with an emphasis on the pixel-level retrievals (referred to as Level-2 products), with 1-km or 5-km spatial resolution at nadir. An example of each Level-2 cloud product from a common data granule (5 min of data) off the coast of South America will be discussed. Future efforts will also be mentioned. Relevant points related to the global gridded statistics products (Level-3) are highlighted though additional details are given in an accompanying paper in this issue.

The MODIS cloud products: algorithms and examples from Terra

Mapping evapotranspiration at high resolution with internalized calibration (METRIC) is a satellite-based image-processing model for calculating evapotranspiration (ET) as a residual of the surface energy balance. METRIC uses as its foundation the pioneering SEBAL energy balance process developed in The Netherlands by Bastiaanssen, where the near-surface temperature gradients are an indexed function of radiometric surface temperature, thereby eliminating the need for absolutely accurate surface temperature and the need for air-temperature measurements. The surface energy balance is internally calibrated using ground-based reference ET to reduce computational biases inherent to remote sensing-based energy balance and to provide congruency with traditional methods for ET. Slope and aspect functions and temperature lapsing are used in applications in mountainous terrain. METRIC algorithms are designed for relatively routine application by trained engineers and other technical professionals who possess a fami...

Satellite-Based Energy Balance for Mapping Evapotranspiration with Internalized Calibration (METRIC)—Model

The Department of Energy's Atmospheric Radiation Measurement (ARM) program sponsored a large intensive observation period (IOP) to study aerosol during the month of May 2003 around the Southern Great Plains (SGP) Climate Research Facility (CRF) in north central Oklahoma. Redundant measurements of aerosol optical properties were made using different techniques at the surface as well as in vertical profile with sensors aboard two aircraft. One of the principal motivations for this experiment was to resolve the disagreement between models and measurements of diffuse horizontal broadband shortwave irradiance at the surface, especially for modest aerosol loading. This paper focuses on using the redundant aerosol and radiation measurements during this IOP to compare direct beam and diffuse horizontal broadband shortwave irradiance measurements and models at the surface for a wide range of aerosol cases that occurred during 30 clear-sky periods on 13 days of May 2003. Models and measurements are compared over a large range of solar-zenith angles. Six different models are used to assess the relative agreement among them and the measurements. Better agreement than previously achieved appears to be the result of better specification of input parameters and better measurements of irradiances than in prior studies. Biases between modeled and measured direct irradiances are less than 1%, and biases between modeled and measured diffuse irradiances are less than 2%.

/pdf/radiative-closure-studies-for-clear-skies-during-the-arm-4pbb4t1lkz.pdf

Radiative Closure Studies for Clear Skies During the ARM 2003 Aerosol Intensive Observation Period

The statistics of natural backgrounds extracted from an Airborne Visible and Infrared Imaging Spectrometer (AVIRIS)
hyperspectral datacube collected over Fort AP Hill, VA, were used to demonstrate the effects of the two atmospheric
components of a statistical end-to-end performance prediction model. New capabilities in MODTRAN TM 5 were used to
generate coefficients for linear transformations used in the atmospheric transmission and compensation components of a
typical end-to-end model. Model radiance statistics, calculated using reflectance data, is found to be similar to the
original AVIRIS radiance data. Moreover, if identical atmospheric conditions are applied in the atmospheric
transmission and in the atmospheric compensation model components and the effects of sensor noise are disregarded, the
resulting reflectance statistics are identical to the original reflectance statistics.

Modeling the effects of atmospheric propagation for spectral libraries of natural backgrounds

Radiance measurements conducted from a high-altitude platform to retrieve surface properties will potentially involve long, near-horizontal viewing geometries. The computer code MODTRAN is widely used for the prediction of the propagation of infrared radiation through the lower atmosphere. Consequently, we have undertaken to test the predictions of MODTRAN for the 3 - 5 and 8 - 12 micron spectral regions under mid-Eastern desert conditions. This paper compares experimental measurements in geometries of interest with calculations using the latest version of MODTRAN. Results indicate a strong dependence of the remotely sensed radiation on both the aerosol and water vapor content.

Comparison of atmospheric transmittance measurements in the 3- to 5- and 8- to 12-μm spectral regions with MODTRAN: considerations for long near-horizontal path geometries

A rapid 3-5 micron tda simulation based on modtran4

Spectral radiance measurements in the solar reflective regime, which are obtained from downward looking space-based or airborne imaging spectrometer systems must be compensated for the influence of the atmosphere in order to retrieve earth surface reflectance spectra. The Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) atmospheric correction code was updated with a new option to handle measurements from a new pushbroom sensor system having a continuously varying viewing geometry over an image scene. However, the new technique employed in FLAASH to handle this atmospheric compensation scenario does not explicitly account for the different viewing geometry values for each line in an image scene, which causes some error in the FLAASH retrievals. Simulated radiance cubes are used to gain a quantitative understanding of FLAASH retrieval errors due to viewing geometry errors.

Gail P. Anderson

Papers

Radiative Closure Studies for Clear Skies During the ARM 2003 Aerosol Intensive Observation Period

Modeling the effects of atmospheric propagation for spectral libraries of natural backgrounds

Comparison of atmospheric transmittance measurements in the 3- to 5- and 8- to 12-μm spectral regions with MODTRAN: considerations for long near-horizontal path geometries

A rapid 3-5 micron tda simulation based on modtran4

Atmospheric compensation for imaging spectrometer systems with changing imaging geometry