Donald W. Deering

A sphere-scanning radiometer for rapid directional measurements of sky and ground radiance

A unique field instrument, called the PARABOLA, a collapsable support boom, which is self contained and easily transportable to remote sites to enable the acquisition of radiance data for almost the complete (4 pi) sky and ground-looking hemispheres in only 11 seconds was designed. The PARABOLA samples in 15 deg instantaneous field of view sectors in three narrow bandpass spectral channels simultaneously. Field measurement on a variety of earth surface cover types using a truck boom, a specially designed pickup truck mounting system, and a hot air balloon were studied. The PARABOLA instrument has potential for climatological and other studies which require characterization of the distribution of diffuse solar radiation within the sky hemisphere.

Estimating spectral albedo and nadir reflectance through inversion of simple BRDF models with AVHRR/MODIS‐like data

In recent years, many computationally efficient bidirectional reflectance models have been developed to account for angular effects in land remote sensing data, particularly those from the NOAA advanced very high resolution radiometer (AVHRR), polarization and directionality of the Earths reflectances (POLDER), and the planned EOS moderate-resolution imaging spectrometer (MODIS) and multi-angle imaging spectroradiometer (MISR) sensors. In this study, we assessed the relative ability of 10 such models to predict commonly used remote sensing products (nadir reflectance and albedo). Specifically, we inverted each model with ground-based data from the portable apparatus for rapid acquisition of bidirectional observations of the land and atmosphere (PARABOLA) arranged in subsets representative of satellite sampling geometries. We used data from nine land cover types, ranging from soil to grassland (First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE)) to forest (Boreal Ecosystem-Atmosphere Study (BOREAS)). Retrieved parameters were used in forward model runs to estimate nadir reflectance and spectral albedo over a wide range of solar angles. We rank the models by the accuracy of the estimated products and find results to be strongly dependent on the view azimuth angle range of the inversion data, and less dependent on the spectral band and land cover type. Overall, the nonlinear model of Rahman et al. [993] and the linear kernel-driven RossThickLiSparse model [Wanner et al., 1995] were most accurate. The latter was at least 25 times faster to invert than the former. Interestingly, we found these two models were not able to match the various bidirectional reflectance distribution function (BRDF) shapes as well as other models, suggesting their superior performance lies in their ability to be more reliably inverted with sparse data sets. These results should be useful to those interested in the computationally fast normalization of bidirectional reflectance data and the estimation of radiation flux parameters (albedo, absorbed radiation) over diverse land covers.

Structure analysis and classification of boreal forests using airborne hyperspectral BRDF data from ASAS

Abstract A new approach is presented for deriving vegetation canopy structural characteristics from hyperspectral bidirectional reflectance distribution function (BRDF) data. The methodology is based on the relationship between spectral variability of BRDF effects and canopy geometry. Tests with data acquired with the Advanced Solid-State Array Spectroradiometer (ASAS) over Canadian boreal forests during the BOREAS campaign show that vegetation structural characteristics can be derived from the spectral variability of BRDF effects. In addition, the incorporation of both BRDF effects and hyperspectral resolution data substantially improve the classification accuracy. Best classification results are obtained when hyperspectral resolution and BRDF data are combined, but the improvement is not consistent for all classes. For example, adding BRDF information to hyperspectral data increases the overall classification accuracy for a six-class fen site from 37.8% to 44.7%. The addition, however, reduces the accuracy for the jack pine class from 43.6% to 28.8%. These new findings provide evidence for improved capabilities for applications of MISR and MODIS data. The spectral resolution of MODIS is expected to be sufficient to derive canopy structural information based on the spectral variability of BRDF effects, and for MISR a significant improvement of classification accuracies can be anticipated from the combination of nadir reflectance and off-nadir data.

Characterization of the Reflectance Anisotropy of Three Boreal Forest Canopies in Spring–Summer

Abstract As a part of the Boreal Ecosystem-Atmosphere Study (BOREAS), measurements of the spectral reflectance anisotropy of three boreal forest canopies were studied for cloudless sky conditions at the phenological growth stages which were at or near maximum leaf area index at each site. The three sites were relatively homogeneous mature stands of black spruce, jack pine, and aspen located in the southern boreal zone of central Saskatchewan. Measurements of the spectal bidirectional reflectance factors with a 15° instrument field of view in three spectral bands centered at 662 nm, 826 nm, and 1658 nm were made with the PARABOLA instrument over a range of solar zenith angles typically varying from 35° (near solar noon) to 70°. The measured reflectance factors showed large anisotropy at all three sites and for all three wavelengths, with prominant backscatter peak reflectances, and strong retro solar view angle (hot spot) maximum reflectances in the visible (662 nm) and shortwave infrared (1658 nm) for the jack pine and black spruce sites, with a less pronounced hot spot at the aspen site. Pronounced effects of canopy and understory shadowing in the visible, as a function of solar zenith angle (SZA), were observed for the black spruce and jack pine sites, with resultant large linear increases in computed normalized difference and simple ratio vegetation indices as SZA increased for near-nadir view angles. Hemispheric spectral reflectances or spectral albedos were computed from angular integration of PARABOLA measured bidirectional reflectances. Visible (662 nm) hemispheric reflectances for the jack pine and black spruce canopies showed very little variation with solar zenith angle, while near-infrared hemispheric reflectances increased strongly with increasing SZA. Estimates were made of the total shortwave albedo for the aspen and jack pine sites from irradiance and reflectance weighting of the spectral hemispheric reflectances in the three measured wavelengths. Comparison of estimated to pyranometer measured total albedo showed all estimates to be biased high, but only by about 0.007–0.018, depending on which of two sets of pyranometer measured albedos were utilized for the comparison. The measured bidirectional reflectance factor (BRF) data sets reported in this study coupled with ancillary data of biophysical parameters collected at the same sites by BOREAS researchers provide a unique data set for the development and characterization of canopy bidirectional reflectance modeling and for the interpretation of remotely sensed data for boreal forest canopies.

Prairie grassland bidirectional reflectances measured by different instruments at the FIFE site

Land surface reflectance measurements were acquired during the First ISLSCP Field Experiment (FIFE) field campaigns using a variety of ground-based and airborne spectral radiometers. To examine the validity of the assumption that the values acquired by the several different instruments and teams were interchangeable, the surface radiation measurement teams converged on a common site for 1 day during the fifth intensive field campaign (IFC 5) in 1989. The instruments compared for near-surface measurements included two ground-based Spectron Engineering SE59Os, one helicopter-mounted SE590, one ground-based and one helicopter-mounted Barnes modular multiband radiometer (MMR), and the portable apparatus for rapid acquisitions of bidirectional observations of land and atmosphere (PARABOLA) field radiometer. Comparisons were made for nadir measurements over a range of solar zenith angles and a range of off-nadir viewing angles. The bidirectional reflectances from the different instruments were generally found to be quite comparable. For example, for a 52° solar zenith angle, the nadir red and near-infrared spectral reflectance factors ranged from 3.5 to 4.5% and 28.2 to 31.9%, respectively. At the smaller solar zenith angles, however, the differences were somewhat greater (red, 4.5–6.1%; near-infrared (NIR), 25.0–28.9%), and the coefficients of variation for the samples taken by all of the instruments increased. Off-nadir viewing caused major departures from nadir bidirectional reflectances (30% reflectance at nadir compared with 55% at 60° off nadir in the NIR, for example), but all of the instruments captured the effects reasonably well. Spectral vegetation indices were found to have a considerable dependence on both solar zenith angle and sensor viewing angle. In spite of the general agreement between most instruments and teams, the lack of a more consistent band-to-band agreement resulted in appreciable differences in the spectral vegetation index values, which could potentially affect the accuracy and precision of remote sensing assessments of biophysical parameters.

Bidirectional reflectances of selected desert surfaces and their three-parameter soil characterization.

Abstract Spectral bidirectional reflectances were measured over three natural alkali soil sites using a specially designed radiometer called the Portable Apparatus for Rapid Acquisition of Bidirectional Observations of Land and Atmosphere (PARABOLA). Two of the sites were void of vegetation with different surface characteristics (alkali flat and dune sand flat), while the third had a sparse cover of desert scrub. The reflectances were strongly non-Lambertian for all three surfaces, but with markedly different patterns. The measured data were fitted with a quasi-physical reflectance model in which the surface backscattering and forwardscattering are separately formulated. A soil reflectance characterization was obtained by assessing the fractional contributions of the forward, backward and Lambertian components. For the desert scrub surface the backscattering component was larger by a factor of 10 or more than the forwardscattering component and by a factor of about two larger than the Lambertian component. The alkali flat backscatter component was little more than half the magnitude of the Lambertian component, and the dune sand was characterized by a forwardscatter component more than double the backscatter component. This three-parameter characterization produced a satisfactory fit to the measured reflectances. Since the component values remain relatively constant at different sun angles, this approach appears very promising as a basis for soil patterns or terrain categorization based on their anisotropic reflectance patterns.

Modeling BRF and Radiation Regime of Boreal and Tropical Forests: I. BRF

Abstract Monitoring of forest evolution and functioning with remote sensing depends on canopy BRF (bidirectional reflectance factor) sensitivity to biophysical parameters and to canopy PAR (photosynthetically active radiation) regime. Here, we study the canopy BRF of a tropical (Sumatra) and three boreal (Canada) forest sites, with the DART (discrete anisotropic radiative transfer) model. The behavior of PAR regime of these forests is analyzed in a companion article. We assessed the BRF sensitivity to some major experimental parameters (scale of analysis, viewing and illumination directions, sky radiation) and compared it with BRF sensitivity to commonly studied biophysical quantities: Leaf area index (LAI) and leaf optical properties. Simulations showed that BRF directional anisotropy is very large for all forests. For example, maximum relative reflectance difference with view zenith angle less than 25° is around 0.5 in the visible, 0.4 in the short wave infrared, and 0.25 in the near-infrared for tropical forest. We showed that this BRF variability associated with experimental conditions can hamper the remote detection of forest LAI and tree cover change such as deforestation of tropical forest. DART BRFs of the boreal sites were favorably compared with ground (PARABOLA) and airborne (POLDER) measured BRFs. This work stressed 1) the potential of the DART model, 2) the importance of accurate field data for validation approaches, and 3) the very strong influence of canopy architecture on forest BRF; for example, depending on forest sites, a LAI increase may imply that nadir near-infrared reflectance increases or decreases.

A simple analytical function for bidirectional reflectance

A simple bidirectional reflectance model based on physical scattering laws is developed to calculate the bidirectional reflectance of a wide variety of surfaces. Chandrasekhars radiative transfer solution, obtained for the anisotropically scattering semi-infinite medium, is used to compute the multiplescattered radiances. However, since the radiative transfer solution does not take architectural effects into consideration, we adopt Hapkes (1986) approach and add an empirical term to explain the hot-spot phenomenon and use Cox and Munks (1954) formulation to take into account the specular reflection. The physical parameters of the model are retrieved from bidirectional reflectance measurements. For each surface, only one set of model parameters is needed for application to all illumination and viewing geometries. The validity of the model is established by comparing the computed and measured reflectances for sets of viewing and illumination angles that were not included in the inversion algorithm. Good agreement is shown between the model-computed and the observed reflectances for dense prairie vegetation canopies, a sparse desert scrub community, a plowed agricultural field, and an alkali flat.

Reflectance anisotropy for a spruce-hemlock forest canopy

Abstract The bidirectional reflectance distribution characteristics of a spruce-hemlock forest were studied during field campaigns conducted in August-September of two different years using the three-channel PARABOLA radiometer system and a SE590 spectrometer mounted for sampling above a forest canopy. Canopy bidirectional reflectance data were acquired under clear, cloudless sky conditions over a large range of solar zenith angles. Pyranometer measurements of the total hemispherical upwelling and downwelling solar radiation above and below the forest canopy were simultaneously acquired. In addition, the PARABOLA instrument was deployed at different heights within the canopy, under overcast sky conditions, to measure the directional downward spectral radiation transmitted. The highly absorbing forest canopy, with a total plant area index of 3.9, reflected only ∼ 2.5% of the red (0.662 μm) hemispherical irradiance. Reflectances for wavebands in the photosynthetically active radiation (0.4–0.7 μm), the near-infrared (NIR, 0.826 μm), and the shortwave infrared (SWIR, 1.658 μm) were approximately 3%, 26%, and 15%, respectively. The bidirectional spectral reflectance factors in the solar principal plane were highly anisotropic, while those in the plane perpendicular to the solar principal plane were much closer to Lambertian. The forwardscatter was not much greater than the nadir reflectance, differing from the usual dense vegetative cover cases previously examined, especially for the NIR waveband. The hot spot maximum reflectance, occurring at the retrosolar view angle, was pronounced in all three wavebands. This prominent hot spot effect has not been observed for the NIR in other vegetation canopy types (e.g., grasslands, agricultural crops, and semiarid and desert scrub). View angle effects were more prominent than solar zenith angle effects, except at large solar zenith angles. Hemispherical spectral reflectance factors and albedo increased with increases in solar zenith angle. Substantial variation was observed in the normalized difference vegetation index (NDVI) for different view and solar zenith angles with a nominal value of 0.75.

Biophysical properties affecting vegetative canopy reflectance and absorbed photosynthetically active radiation at the FIFE site

Leaves of the dominant grass species of the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) site reflect and transmit radiation in a similar manner to other healthy green leaves. Visible reflectance factors (RFs) and transmittance factors (TFs) were lower for older leaves than younger leaves except during senescence, when RF and TF values were higher. Near-infrared (NIR) RF values increased and TF values decreased with leaf age, with the reverse occurring as the leaf underwent senescence. Leaf optical properties were not found to be dependent on leaf water potential in the range from −0.5 to −3.0 MPa. Canopy bidirectional reflectance factor (BRF) values generally increased with increasing view zenith angle (θυ). Maximum values were in the backscatter direction, whereas BRF values in the visible region were lowest at oblique off-nadir θυ in the forward scatter direction and at or near nadir in the NIR region. Solar principal plane BRF values varied most at large solar zenith angles (θs). Visible and mid-infrared canopy BRF values decreased and NIR BRF values increased with leaf area index (LAI). Soil BRF distributions in the solar principal plane varied slightly with θs and θυ and varied considerably for wet and dry surfaces. Spectral vegetation indices (SVIs) varied with θs and θυ; values were lowest in the backscatter direction and highest in the forward scatter direction. The fraction of absorbed photosynthetically active radiation (APAR) increased with increasing θs. APAR had a strong linear relationship to nadir-derived SVI values but not to oblique off-nadir-derived SVI values. The relatively small dependence of off-nadir SVI values on θs should allow daily APAR values to be estimated from measurements made at any time of the day.