George Alan Blackburn

Quantifying Chlorophylls and Caroteniods at Leaf and Canopy Scales: An Evaluation of Some Hyperspectral Approaches

An investigation was undertaken into the effectiveness of a range of hyperspectral approaches for estimating the concentrations of chlorophyll a, chlorophyll b and the carotenoids at the plant canopy and leaf scales. Measurements of spectral reflectance, biophysical characteristics and pigment composition were made for a bracken (Pteridium aquilinum) canopy, and its components, throughout a growing season. The results indicate that narrow-band reflectance indices, such as the pigment-specific simple ratio (PSSR), can be developed which have extremely strong relationships with the concentration per unit area of individual pigments at the canopy scale. Spectral derivative approaches, particularly those based on pseudo absorbance (Log 1/R), are also closely related to canopy pigment concentration per unit area, but are more useful for deriving estimates of the concentration per unit mass of photosynthetic pigments at both canopy and leaf scales.

Spectral indices for estimating photosynthetic pigment concentrations : A test using senescent tree leaves

Abstract The possibility of estimating the concentration of individual photosynthetic pigments within vegetation from reflectance spectra offers great promise for the use of remote sensing to assess physiological status, species type and productivity. This study evaluates a number of spectral indices for estimating pigment concentrations at the leaf scale, using samples from deciduous trees at various stages of senescence. Two new indices (PSSR and PSND) are developed which have advantages over previous techniques. The optimal individual wavebands for pigment estimation are identified empirically as 680nm for chlorophyll a, 635nm for chlorophyll b and 470nm for the carotenoids. These wavebands are justified theoretically and are shown to improve the performance of many of the spectral indices tested. Strong predictive models are demonstrated for chlorophyll a and b, but not for the carotenoids and the paper explores the reasons for this.

A real-time hydrological model for flood prediction using GIS and the WWW

The purpose of this paper is to examine the current status of real time hydrological models used for flood nowcasting and hazard mitigation and indicate how WWW-based systems can overcome some of the limitations of existing systems. Whilst hydrologically innovative and robust models are available, they are poorly suited to real time application, are often not well integrated with spatial datasets such as GIS. Current systems also lack flexibility, customisability and accessibility by a range of end users. We describe the development of a Web-based hydrological modelling system that permits integrated handling of real-time rainfall data from a wireless monitoring network. A spatially distributed GIS-based model is integrated on the basis of this incoming data, approximating real-time to produce data on catchment hydrology and runoff. The data can be accessed from any WWW interface, and they can be analysed online using a number of GIS and numerical functions. We discuss the potential users of such a system and the requirements for interfacing model output with these users for hydrological nowcasting and spatial real-time, emergency decision support. Rather than discuss developments in the modelling of hydrology for flood hazard mitigation, this paper focuses on developments in interfacing these models with end users.

Relationships between Spectral Reflectance and Pigment Concentrations in Stacks of Deciduous Broadleaves.

This article presents the findings of a laboratory experiment using stacks of leaves obtained from four species of deciduous tree at various stages of senescence. This methodology generated a wide variation in pigment concentrations (per unit leaf area and per unit ground area) and leaf area index, as may be expected at the plant canopy scale, and provided independent variation between the chlorophylls and carotenoids and leaf area index. Spectral reflectance was measured for each leaf increment of each stack using a spectroradiometer and these data were used to evaluate the nature and strength of relationships between a number of existing and novel spectral transformations and the concentrations of pigments (per unit ground area) within the leaf stacks. Reflectance in narrow wavebands within the visible region was moderately related to chlorophyll concentrations, while ratio indices, employing wavebands in the near-infrared and visible, particularly the green, were more successful. While the wavelength position of the red edge was related to chlorophyll concentration, characteristics of the amplitude of the first and second derivatives of reflectance and pseudo absorbance were more strongly correlated with chlorophyll. No transformation of spectral reflectance that was tested was strongly related to the concentration of carotenoids but two ratio indices were highly correlated with the ratio of carotenoids to chlorophyll a. Interestingly, leaf area index (LAI) was unrelated to all spectral ratio indices including the broad-band normalised-difference and the simple ratio vegetation indices. This indicates that spectral indices may have limited applicability for estimating the LAI of vegetation canopies when the leaves that comprise these canopies have differing chlorophyll concentrations (per unit leaf area).

Mapping individual tree location, height and species in broadleaved deciduous forest using airborne LIDAR and multi-spectral remotely sensed data.

Automated feature extraction based on prototypes is only partially successful when applied to remotely sensed imagery of natural scenes due to the complexity and unpredictability of the shape and geometry of natural features. Here, a new method is developed for extracting the locations of treetops by applying GIS (Geographical Information System) overlay techniques and morphological functions to high spatial resolution airborne imagery. This method is based on the geometrical and spatial properties of tree crowns. Airborne data of the study site in the New Forest, UK included colour aerial photographs, LIDAR (Light Detection And Ranging) and ATM (Airborne Thematic Mapper) imagery. A DEM (Digital Elevation Model) was generated from LIDAR data and then subtracted from the original LIDAR image to create a Canopy Height Model (CHM). A set of procedures using image contouring and the manipulation of the resulting polygons was implemented to extract treetops from the aerial photographs and the CHM. Criteria were developed and threshold values were set using a supervised approach for the acceptance or rejection of features based on field knowledge. Tree species were mapped by classifying the ATM data and these data were co‐registered with the treetop layer. For broadleaved deciduous plantations the success of treetop extraction using aerial photographs was 91%, but was much lower using LIDAR data. For semi‐natural forests, the LIDAR produced better results than the aerial photographs with a success of 80%, which was considered high, given the complexity of these uneven aged stands. The methodology presented here is easy to apply as it is implemented within a GIS and the final product is an accurate map with information about the location, height and species of each tree.

Remote sensing of forest pigments using airborne imaging spectrometer and LIDAR imagery

Abstract This study created and tested predictive models developed using airborne imaging spectrometer and light detection and ranging (LIDAR) instruments for estimating the concentrations of photosynthetic pigments in broad-leaved and coniferous forest plantations. Data were acquired using a Compact Airborne Spectrographic Imager (CASI) and an Airborne Laser Terrain Mapping (ALTM) 1020 instrument in midsummer for study sites in the New Forest, England, along with concomitant in situ measurements of canopy properties. The stands used displayed a wide variation in the biophysical and biochemical properties of interest. When employing the imaging spectrometer data alone, there were no relationships between any spectral variables (band reflectance, band ratios, or first derivatives of reflectance) and canopy biophysical and biochemical properties when both broad-leaved and coniferous stands were analysed as a combined data set. However, for the broad-leaved stands alone, curvilinear relationships were found between the wavelength position of the red edge ( λ RE ) and pigment concentrations per unit ground area (e.g., R 2 =0.88** for chlorophyll a [Chl a ]) and per unit leaf mass (e.g., R 2 =0.76** for Chl a ). The predictive value of these models was somewhat limited; for example, the root mean squared error (RMSE) was 300 mg m −2 (27% of the mean) for Chl a concentration per unit ground area and 1.17 mg g −1 (24% of the mean) for Chl a concentration per unit leaf mass. A ratio of a near-infrared and a green band (865 nm/553 nm) was linearly related to leaf area index (LAI) of the broad-leaved stands ( R 2 =0.71**) and the regression model was a reasonable predictor of the LAI for the independent test sites (RMSE=0.88; 18.6% of the mean). Canopy height information derived from the ALTM data was used to mask out canopy gap areas from the CASI imagery of each stand. This process had limited impact on the relationships between spectral and canopy variables for the broad-leaved stands, and λ RE remained unrelated to pigment concentrations per unit ground area for the coniferous stands. However, the masking process substantially improved the strength of the relationship between λ RE and pigment concentrations per unit leaf mass for the coniferous stands (e.g., for Chl a R 2 =0.85**; RMSE of prediction=0.84 mg g −1 [22% of the mean]). Therefore, the study demonstrates that for broad-leaved stands, spectral models can be applied to imaging spectrometer data to quantify forest pigments and LAI with moderate accuracy. For coniferous stands, the use of LIDAR data to remove canopy gap areas from the CASI imagery considerably increases the accuracy of spectral predictive models for quantifying pigment concentrations per unit leaf mass.

Wavelet decomposition of hyperspectral data: a novel approach to quantifying pigment concentrations in vegetation

This paper reports a new approach for quantifying vegetation pigment concentrations through wavelet decomposition of hyperspectral remotely sensed data. Wavelets are a group of functions that vary in complexity and mathematical properties, that are used to dissect data into different frequency components and then characterize each component with a resolution appropriate to its scale. Wavelet analysis of a reflectance spectrum is performed by scaling and shifting the wavelet function to produce wavelet coefficients that are assigned to different frequency components. By selecting appropriate wavelet coefficients, a spectral model can be established between the coefficients and biochemical concentrations. Hence, wavelet analysis has the potential to capture much more of the information contained within high‐resolution spectra than previous approaches and offers the prospect of developing robust, generic methods for pigment determinations. The capabilities of the wavelet‐based technique were examined using reflectance spectra and pigment data collected for a range of plant species at leaf and canopy scales. For the combined data set and all of the individual vegetation types, methods based on wavelet decomposition appreciably outperformed narrowband spectral indices and stepwise selection of narrowband reflectance. However, there was variation between vegetation types in the relative performance of the three different feature extraction techniques employed for selecting the wavelet coefficients for use in predictive models. There was also considerable variability in the performance of predictive models according to the wavelet function used for spectral decomposition and the optimum wavelet functions differed between vegetation types and between individual pigments within the same vegetation type. The research indicates that wavelet analysis holds promise for the accurate determination of chlorophyll a and b and the carotenoids, but further work is needed to refine the approach.

Quantifying the spatial properties of forest canopy gaps using LiDAR imagery and GIS

The spatial properties of gaps have an important influence upon the regeneration dynamics and species composition of forests. However, such properties can be difficult to quantify over large spatial areas using field measurements. This research considers how we conceptualize and define forest canopy gaps from a remote sensing point of view and highlights the inadequacies of passive optical remotely sensed data for delineating gaps. The study employs the analytical functions of a geographical information system to extract gap spatial characteristics from imagery acquired by an active remote sensing device, an airborne light detection and ranging instrument (LiDAR). These techniques were applied to an area of semi-natural broadleaved deciduous forest, in order to map gap size, shape complexity, vegetation height diversity and gap connectivity. A vegetation cover map derived from imagery from an airborne multispectral scanner was used in combination with the LiDAR data to characterize the dominant vegetation types within gaps. Although the quantification of these gap characteristics alone is insufficient to provide conclusive evidence on specific processes, the paper demonstrates how such information can be indicative of the general status of a forest and can provide new perspectives and possibilities or further ecological research and forest monitoring activities.

Mapping crop phenology using NDVI time-series derived from HJ-1 A/B data

With the availability of high frequent satellite data, crop phenology could be accurately mapped using time-series remote sensing data. Vegetation index time-series data derived from AVHRR, MODIS, and SPOT-VEGETATION images usually have coarse spatial resolution. Mapping crop phenology parameters using higher spatial resolution images (e.g., Landsat TM-like) is unprecedented. Recently launched HJ-1 A/B CCD sensors boarded on China Environment Satellite provided a feasible and ideal data source for the construction of high spatio-temporal resolution vegetation index time-series. This paper presented a comprehensive method to construct NDVI time-series dataset derived from HJ-1 A/B CCD and demonstrated its application in cropland areas. The procedures of time-series data construction included image preprocessing, signal filtering, and interpolation for daily NDVI images then the NDVI time-series could present a smooth and complete phenological cycle. To demonstrate its application, TIMESAT program was employed to extract phenology parameters of crop lands located in Guanzhong Plain, China. The small-scale test showed that the crop season start/end derived from HJ-1 A/B NDVI time-series was comparable with local agro-metrological observation. The methodology for reconstructing time-series remote sensing data had been proved feasible, though forgoing researches will improve this a lot in mapping crop phenology. Last but not least, further studies should be focused on field-data collection, smoothing method and phenology definitions using time-series remote sensing data.

Remote sensing of mangrove biophysical properties: Evidence from a laboratory simulation of the possible effects of background variation on spectral vegetation indices

A physical scale model of a mangrove canopy over different backgrounds was used in the laboratory to investigate the relationships between a number of spectral vegetation indices and LAI and percent canopy cover and the sensitivity of these indices to variations in background reflectance properties. High spectral resolution reflectance data were acquired from the experimental canopy and these were used to simulate the response in the red and NIR wavebands of the Landsat TM sensor. These data were then used to calculate NDVI, RVI, DVI, PVI, SAVI, SAVI 2, and TSAVI. Three derivative-based indices (1DL_DGVI, 1DZ_DGVI and 2DZ_DGVI) that measure the amplitude of the chlorophyll red edge were also calculated. Based on the correlation coefficients for both LAI and percent canopy cover, the effects of background variations were most pronounced for NDVI, SAVI and TSAVI, whereas SAVI 2 and RVI were moderately affected. The least affected spectral indices were DVI, PVI, 1DL_DGVI, 1DZ_DGVI and 2DZ_DGVI. The DVI appears to be the optimal spectral vegetation index for estimating the biophysical properties of mangroves which have variable background conditions because it had robust linear relationships with LAI and percent cover and it can be easily derived from commonly available broad band remotely sensed data.