Quantitative assessment of vegetation structural attributes from aerial photography

Cover of vegetation understorey and overstorey was determined from aerial photography at 1:25 000 and 1:40 000 scales by a grid sampling technique. Models were developed relating values of aerial cover to field cover as determined by intensive field measurement. The influence of photo-scale, photo colour, the angle of the image, shadow, the hiatus between aerial and field sampling, crown width, crown height, proportion of dead trees, drought prior to aerial sampling, land type, previously cleared vegetation and incline on explanatory models was also examined. The only variables that could be clearly interpreted as influencing the models were vegetation height, photo-scale and land type. Only the latter two variables are useful for predictive models. The smaller the scale of photography the greater the exaggeration of the aerial image of tree crowns. This probable result of photo graininess would be most significant when tree crowns are small, an inverse surrogate of tree height. Two-phase models were developed for predicting basal area and biomass from aerial cover. In most instances models were successful for predicting overstorey and understorey cover and for predicting total basal area and biomass. The technique offers a powerful and cost-effective method of assessing vegetation change over long time periods in a way that no other technique can duplicate.     

Forest mapping by geoinformatics for landscape fire behaviour modelling in coastal forests,Greece

This study aims at quantifying and mapping fire-related characteristics of forest structure through field inventories, statistics, remote sensing, and geographical information systems in the island of Lesvos, northeast Aegean Sea, Greece. Simulation of fire behaviour requires forest biomass inputs that describe surface fuel types/models along with canopy fuel properties, such as canopy cover, stand height, crown base height, and crown bulk density, to accurately predict surface and crown fire spread and spotting potential. Forest canopy characteristics and other vegetation attributes were sampled and derived in over 100 field plots, the majority of which were located in coastal pine forest stands. Regression models involving four dependent forest stand variables (stand height, canopy cover, crown base height, and crown bulk density) were developed using generalized additive models. The values of adjusted R ² were 0.72 for stand height, 0.68 for canopy cover, 0.51 for crown base height, and 0.33 for crown bulk density. These regression models were used to create forest fuel characteristics layers, which can be used as inputs to fire management applications and state-of-the-art landscape-scale fire behaviour models.     

Vegetation spatial variability and its effect on vegetation indices

Abstract

Landsat MSS data were used to simulate low resolution satellite data, such as NOAA AVHRR, to quantify the fractional vegetation cover within a pixel and relate the fractional cover to the normalized difference vegetation index (NDVI) and the simple ratio (SR). The MSS data were converted to radiances from which the NDVI and SR values for the simulated pixels were determined. Each simulated pixel was divided into clusters using an unsupervised classification programme. Spatial and spectral analysis provided a means of combining clusters representing similar surface characteristics into vegetated and non-vegetated areas. Analysis showed an average error of 12·7 per cent in determining these areas. NDVI values less than 0·3 represented fractional vegetated areas of 5 per cent or less, while a value of 0·7 or higher represented fractional vegetated areas greater than 80 per cent. Regression analysis showed a strong linear relation between fractional vegetation area and the NDVI and SR values; correlation values were 0·89 and 0·95 respectively. The range of NDVI values calculated from the MSS data agrees well with field studies.     

Testing of a model for the simulation of the volume reflectance of water due to suspended sediment under controlled conditions,for various sediment types

Abstract

A model that simulates the volume reflectance of water is tested in laboratory conditions with three different sediments of known grain size distribution and spectral reflectance. The model uses, as its main input variables, the grain size distribution of the suspended sediment, the density of the grains and the spectral reflectance of the wet sediment from which the suspension comes. The spectral bands of the model are completely programmable and they have been adapted to correspond to green, red and near-infrared, as in Landsat MSS bands 4, 5 and 6. The grain size distribution, the reflectance in the three spectral bands and the density of the three sediment types had been measured and these measurements were used as input variables for the model to simulate the volume reflectance of the water, in the same spectral bands, for different suspended sediment concentrations. Each sediment type was used to produce several concentration levels in a specially prepared 1101 tank and the volume reflectance of the water was measured. The measurements were compared with the simulated results of the model. Then the model was modified to simulate the volume reflectance of a stratified environment, where the concentration changes with depth, and used to explain some discrepancies in the chromaticity transformation.     

Spectral separability of tropical forest cover classes

Abstract

Spectral separability analysis of various tropical forest cover classes as recorded on Landsat MSS data were carried out for two test areas of northeastern India. The results indicated that two density classes of forests and two edaphic forest types formed spectrally separable classes but some of the important physiognomic units could not be reliably separated using Landsat MSS data in a spectrally complex tropical forest environment.     

Mapping forest biomass through digital processing of IRS-IA data

The present study deals with the mapping of forest basal cover and biomass using IRS data. IRS-LISS-I data were classified into forest types and crown cover categories. A stand biomass was computed for selected sites using density, basal cover data and biomass estimation equations. Allometric relations were developed between crown cover and basal cover and between crown cover and biomass. Using these relations basal cover and biomass were computed for each crown cover class of each forest type. The classes having identical biomass were merged together. Total biomass for each forest type was computed by using mean values and the aerial extent. The average total above-ground biomass density between forest types ranged between 52–36tha^-1 ∥Plantations) and 371–08tha^-1 (Sal forest). The estimates of the study compared well with the estimates for 19 sites computed through conventional techniques. The method described in the present study is expected to play a significant role in global biomass estimations.     

Spectral texture for improved class discrimination in complex terrain

Abstract

A spatial co-occurrence algorithm has been used to derive image texture from Landsat Multispectral Scanner (MSS) data to increase classification accuracy in a moderate relief, boreal environment in eastern Canada. The aim was to investigate ‘data-driven improvements’, including those available through digital elevation modelling. Overall classification accuracy using MSS data alone was 59·1 per cent when compared to a biophysical inventory of the area compiled primarily by aerial photointerpretation. This increased to 66·2 per cent with MSS plus texture and to 89·8 per cent when MSS data were analysed with geomorphometry extracted from a digital elevation model (DEM). The introduction of MSS texture resulted in statistically significant increases in individual class accuracies in classes that were also well defined using the geomorphometric and integrated data sets. This suggested that some of the additional information provided by geomorphometry was also contained in spectral texture. It was also noted that individual texture orientations resulted in higher class accuracies than average texture measures; this is probably related to structural (slope/aspect) characteristics of specific vegetation communities.     

Satellite remote sensing of rangelands in Botswana I. Landsat MSS and herbaceous vegetation

Field measurements of the cover and biomass of live and dead herbaceous vegetation, the cover of trees and shrubs and the area of bare ground were made for rangelands in three study sites in eastern Botswana between September 1983 and April 1984. The sites were selected to be representative of Terminalia sericea, Cotophospermum mopane and Combretum apiculatum-Acacia nigrescens woodland savannas, which, taken together, occupy a large part of eastern and northern Botswana. Mean herbaceous biomass varied from 0 to 563?kg ha^?1, cover from 0 to 21 per cent and bare ground from 57 to 85 per cent. The mean tree canopy cover in each site was approximately 30 per cent, with a range of 0-50. Landsat miiltispectral scanner (MSS) data were obtained for May, August, November 1983, January and March-April 1984. Nine MSS pixels were registered with 20 sample plots in each site and the ratios of mean band-7 to band-5 digital numbers were calculated. The variation in these ratios between the three sites and four dates on which the data were acquired was analyzed with respect to the field measurements.

The results indicate that the biomass and cover of live herbaceous vegetation and the bare ground individually account for quite small, but significant proportions of the variation in band ratio for all four observation dates taken together. However, when factors that specified site and date were included in the multiple regression models, 75·7, 77·9 and 64·1 per cent of the variation in herb biomass, cover and bare ground respectively were accounted for. Multitemporal integration of the band ratios accounted for 70·3 per cent of the variation in the end-of-season biomass of herbaceous vegetation, without the need to use a site factor. These highly significant relationships were achieved without including measurements of the canopy cover of trees in the models. Analysis of the individual site data revealed some negative relationships between band ratios and both biomass and Cover of herbaceous vegetation in one site, which seem to be a result of a strong negative relationship between the cover of herbaceous vegetation and trees in this vegetation type.

It was concluded that predictive equations could be constructed which enable cover and biomass of herbaceous vegetation and the area of bare ground to be estimated from Landsat MSS band-7/band-5 ratios, but only if the relationship is applied to sites having the same type of vegetation as that for which the equations were derived. Stratification of the scene using vegetation maps is therefore an important requirement for the application of remote sensing by Landsat MSS to the monitoring of the rangelands in Botswana     

Urban built-up land change detection with road density and spectral information from multi-temporal Landsat TM data

In this article, Landsat TM images acquired during the same season from both 1984 and 1997 were analysed for urban built-up land change detection in Beijing, China, where great changes have taken place during the recent decades. To reduce the spectral confusion between urban 'built-up' and rural 'non built-up' land cover categories, we propose a new structural method based on road density combined with spectral bands for change detection. The road density represents one type of structural information while the multiple Landsat TM bands represent spectral information. Road density maps for both dates were produced using a gradient direction profile analysis (GDPA) algorithm and then integrated with spectral bands. Results from the spectral-structural postclassification comparison (SSPCC) and spectral-structural image differencing (SSID) methods were evaluated and compared with spectral-only change detection methods. The proposed SSPCC method greatly reduced spectral confusion and increased the accuracy of land cover classification compared with spectral classification, which in turn improved the change detection results. This article also shows that the SSID change detection result complemented spectral band differencing by detecting areas with greater structural changes, some of which were missed, by spectral band differencing.     

Utility of AVHRR data for land degradation assessment: a case study

AVHRR data are widely used to monitor vegetation greenness and to provide a gross measure of primary production throughout the world. This paper examines whether AVHRR data can be used for determining the extent of land degradation in arid rangelands under commercial grazing using models of vegetation dynamics and animal grazing behaviour developed for Landsat-MSS data. These models are applied after large rainfall events and either search for systematic change in average vegetation cover across relatively uniform landscapes with increasing distance from stock watering points or examine the magnitude of vegetation response to rainfall for each pixel.

We applied the models where previous work with Landsat-MSS had demonstrated the extent of grazing impact. An index of vegetation cover using adjusted AVHRR channel 1 values produced trends in wet period average vegetation cover with increasing distance from water similar to, but less pronounced than, those obtained with MSS data. NDVI produced inconsistent and often ambiguous results when compared with the MSS data. AVHRR-derived vegetation indices were unusable in degradation assessment procedures which require pixel-scale vegetation response models. The large AVHRR pixel, even in LAC mode, creates difficulties in detecting grazing impact. Landscape changes as a result of grazing occur at a finer scale and are therefore subsumed within the pixel. Misregistration of multi-temporal images further reduces the ability to detect grazing impact on a pixel basis when such change is occurring within the pixel.

We conclude that despite their cost attractiveness, AVHRR data are inappropriate for the reliable detection of grazing impact using grazing gradient methods in the large paddocks of arid rangelands.     

Discrimination of conifer height,age and crown closure classes using Landsat-5 TM imagery in the Canadian Northwest Territories

Empirical relationships between forest stand attributes and Landsat-5 Thematic Mapper spectral response were developed in order to assess its informational value in support of forest inventory operations in the Northwest Territories. An existing large-area classification procedure, based on a supervised methodology, has been able to generate classes of white spruce and jack pine. Within individual forest species groups, spectral variability related to differences in crown closure, height and age is of interest to forest managers in the region. The objective of this study was to determine the accuracy with which stands within the white spruce and jack pine classes could be further separated into two stand height classes (< 15 m and S 15 m), two age classes (< 100 years and S 100 years) and two crown closure classes (< 30% and S 30%) with a single (summer) Landsat-5 Thematic Mapper (TM) image. Discrimination generally improved with the addition of spectral texture measures where independently assessed accuracies ranged from 60 to 90%. A look-up table was devised for conifer-dominated areas (> 80% dominant species) which could subsequently be assigned for height, age and crown closure class values based on Landsat TM spectral response patterns.     

Satellite remote sensing of rangelands in Botswana II. NOAA AVHRR and herbaceous vegetation

NOAA-7 Advanced Very High Resolution Radiometer (AVHRR) global-area coverage (GAC) data for the visible and near-infrared bands were used to investigate the relationship between the normalized difference vegetation index (NDVI) and the herbaceous vegetation in three representative rangeland types in eastern Botswana. Regressions between Landsat MSS band-7/band-5 ratios and field measurements of the cover of the live parts of herbaceous plants, above-ground biomass of live herbaceous plants and bare ground were used in conjunction with MSS data in order to interpolate the field data to 144 km² areas for comparison with blocks of nine AVHRR GAC pixels. NOAA NDVI data were formed into 10-day composites in order to remove cloud cover and extreme off-nadir viewing angles. Both individual NDVI composite data and multitemporal integrations throughout the period May 1983-April 1984 were compared with the field data.

In multiple linear regressions, the cover and biomass of live herbaceous plants and bare ground measurements accounted for 42, 56 and 19 per cent respectively of the variation in NDVI. When factors were included in I he regression models to specify the site and date of acquisition of the data, between 93 and 99 per cent of the variation in NDVI was accounted for. The total herbaceous biomass at the end of the season was positively related to integrated NDVI, up lo the maximum biomass observed in a 12km × 12km area (590kgha^?1)- These results give a different regression of herbaceous biomass values on integrated AVHRR NDVI to that reported by Tucker et at. (1985 b) for Senegalese grasslands. The effect of the higher cover of the tree canopy in Botswana on this relationship and on the detection of forage available to livestock is discussed.     

Production of land-use and land-cover maps of central Guangdong Province of China from LANDSAT MSS imagery†

Abstract

The problems of mapping the land use and land cover of a large area of central Guangdong Province of China from LANDSAT MSS data were examined with reference to the manual and digital approaches. Based on intensive field study of a test site in the study area, we discovered the importance of topographic effects, slopes, seasons, cropping system and the intensity of land use in affecting the accuracy of the resultant maps. It was concluded that visual interpretation was essential in providing the level of reference required for the image interpreter to perform the digital analysis satisfactorily. In view of the coarse spatial resolution of the LANDSAT MSS data it is recommended that the most straightforward digital analysis involving the use of the supervised approach with the minimum Euclidean distance classification and an iterative selection of training areas be adopted for the land-use/land-cover mapping of the whole study area to achieve an accuracy of 80 per cent for eight level I and fifteen level II categories of the land-use and land-cover classification scheme.     

Pixel-mixing effects and their significance to identifying snow condition from LANDSAT MSS data

Abstract

Radiometric measurements on snow in LANDSAT MSS wavebands have shown a correlation between snow condition and the ratio of green/infrared (band 4/band 7) wavebands. Systematic changes from wet dense snow at tow altitudes to dry less dense snow at higher altitudes should be revealed by a decrease in the intensity ratio band 4/band 7 with altitude. However, analysis of spring LANDSAT MSS images for the Cairngorm Mountains shows that the intensity ratio band 4/band 7 actually increases with altitude. A mixed-pixel hypothesis is invoked to account for this pattern. The results suggest that only in areas where the snow cover is continuous can information on snow condition be reliably obtained from LANDSAT MSS data.     

The causes of age dependent changes in the spectral response of lavas,Craters of the Moon,Idaho, U.S.A

The Craters of the Moon basaltic lavas erupted between about 15000 and 2000 years ago in short periods occurring about every 2000 years. On LANDSAT MSS imagery the appearance of pahoehoe flows with ‘Blue Dragon’ crusts, which has previously been used as an aid to mapping and correlation across the lava field, changes systematically with age. The overall brightness and the slope of the reflectance curve between band 4 and band 7 both increase with age. In situ measurements of the bidirectional reflectance factor (BRF) of natural rock surfaces in the area shows that this ageing of the remotely sensed spectral response is a combination of (1) a change in the spectral response of the rock surface due lo weathering, (2) growth of lichen and moss over the rock surface, both factors causing little further change after about 8000 years, and (3) settling of windblown sand and establishment of vascular plants (grasses and shrubs), which continues with age. The relative importance of the last factor is not directly assessable from ground based radiometer measurements; however both weathering and lichen produce similar, complementary changes in BRF which can account for much of the spectral evolution seen on MSS images.     

Integration of Landsat and aeromagnetic data as aid to the structural analysis of Crete and S.E. Peloponessus

Abstract

The paper comments on the usefulness of remotely-sensed data (Land-sat MSS images in both digital and photographic format—aeromagnetic data) in the tectonic analysis of areas of Greece. The island of Crete was the main case study area, while a general analysis was also carried out in the South Eastern Peloponessus.

Both image processing (spectral and spatial analysis) of the Landsat CCTs of Crete and computer analysis of the features mapped on the images have been carried out. Aeromagnetic data are also analysed using advanced processing techniques (spectral analysis and deconvolution). Structural interpretations are improved by the study of the enhanced Landsat images and aeromagnetic maps, while the use of the various computer techniques makes the analysis of the mapped patterns easier and more accurate.

The combined interpretation of aeromagnetic and Landsat MSS data added several significant structural features, previously unrecognised from separate interpretations of aeromagnetic data and Landsat images.     

Shuttle radar mapping with diverse incidence angles in the rainforest of Borneo

Abstract

Shuttle Imaging Radar-B (SIR-B) images, obtained at two different incidence angles were analysed for discrimination and mapping of vegetation in the rainforest of Borneo. In an area of coastal lowland three units of forest canopy and two units of open surface cover were distinguished and mapped. The backscatter characteristics of the units were compared and analysed by ratioing digital image data of sample sites against digital data of reference sites. The ratios show that the backscatter for each surface unit is clustered in separate, mostly non-overlapping, domains at both incidence angles. Different rates of change with incidence angle indicate a strong angular dependence in the dominant backscatter mechanism of swamp that is not apparent for the other units. A corresponding digitally coregistered Landsat multispectral scanner (MSS) image helped to verify the spatial distribution of the mapped units in the coastal lowland. In the interior upland three major forest species associations that have contrasted canopy structures were not discriminated on the SIR-B images. Owing to perennial cloud cover the distribution and extent of the different canopies cannot be determined from Landsat MSS or other images obtained at optical wavelengths. The ability to discriminate and map different forest canopies in the interior of Borneo requires a wider radar response capability which may be obtainable at shorter wavelengths and with multipolariz-ation states.     

Radiative transfer modelling of discontinuous tree canopies at microwave frequencies

Abstract

Development of a first-order radiative transfer model for predicting backscatter from tree canopies has been underway at the University of Michigan Radiation Laboratory for some time. This model is known as the Michigan Microwave Canopy Scattering (MIMICS) model. The first-generation model, MIMICS I, was developed for canopies with continuous (closed) crown layers and its validity has been verified in several modelling analyses. This article presents the second-generation MIMICS model (MIMICS II) which accounts for canopies with discontinuous (open) crown layer geometries. MIMICS II models open crown layers by treating the location, size and shape of the individual tree crowns as random variables. The backscattering coefficients for the canopy are then determined by introducing statistics derived from these parameters into the radiative transfer solution.

Application of the radiative transfer equations to the discontinuous canopy geometry is presented. The application of random variables defining the crown geometry and the incorporation of these variables into the radiative transfer solution is discussed. The resulting model is a robust fully polarimetric solution that is applicable over a wide variety of canopy architectures. Model simulations are compared to results generated with the continuous canopy model. The effect of the open crown geometry is found to be most significant at shallow incidence angles and at high frequencies for trees with well-developed crowns. Under these conditions, the gaps in the crown layer give rise to a notable increase in crown layer transmissivity which allows the radar to see through to the lower layers of the canopy more easily thereby directly affecting the backscatter contribution of the trunks and ground.     

Detection of forest change in the Green Mountains of Vermont using Multispectral Scanner data

Abstract

This study evaluates the potential of Landsat Multispectral Scanner (MSS) data to monitor long term changes in high-elevation coniferous forests. August 1973 and 1984 MSS data from the Green Mountains of Vermont were used. Following co-registration and standardization of data sets, colour density slices of 0.75 μm and 0·95 μm band difference images were produced. False colour composites using either (a) the 1973 0·65 μm and 0·75 μm band in conjunction with the 0.75 μm band difference data set or (b) the 1973 0·65 μm and 0·95 μm band in conjunction with the 0.95 μm band difference data set, were also produced. These images indicated that major reflectance decreases occurred for the near-infrared bands between 1973 and 1984 for much of the high-elevation forests. This includes the west-facing slope of Camels Hump, an area for which long term records documenting decreases in basal area and inferred biomass are available. It is believed that the reflectance decreases noted for the high-elevation coniferous forests are related to the forest decline process and are associated with reduction in green leaf biomass and/or the increased levels of mortality and higher amounts of dead branches accompanying this reduction in biomass.     

The relation between reflected radiation and yield on the Broadbalk winter wheat experiment†

Abstract

In the classical experiment on Broadbalk field, Rothamsted, winter wheat has been grown continuously under various treatments since 1843. Reflected radiation in red (R) and near-infrared (NIR) wavebands was measured over the field in 1987, twice with an airborne multispectral scanner (MSS) from an altitude of 600m and five times with a muhiband radiometer from 2 m above the crop surface. The normalized difference vegetation index, NDVI= ( NIR?R)/ (NIR + R), was calculated for each date and its relation with harvest yield investigated.

The NDVI determined from the airborne MSS data was correlated with yield, and the correlation was found to increase if the variable effect of productivity was introduced into the relationship. The NDVI values calculated from the ground radiometry were more strongly correlated with yield, however. The differences in yield on Broadbalk are caused mainly by the amounts of nitrogen-containing fertilizer applied. The results suggest that the radiation measured by airborne MSS can give a rough guide to yield and nitrogen nutrition, whereas ground radiometry could be used to predict yield and potential deficiencies of nitrogen.