Evaluating the potential of multispectral imagery to map multiple stages of tree mortality

Insect outbreaks are major forest disturbances, causing tree mortality across millions of ha in North America. Resultant spatial and temporal patterns of tree mortality can profoundly affect ecosystem structure and function. In this study, we evaluated the classification accuracy of multispectral imagery at different spatial resolutions. We used four-band digital aerial imagery (30-cm spatial resolution and aggregated to coarser resolutions) acquired over lodgepole pine-dominated stands in central Colorado recently attacked by mountain pine beetle. Classes of interest included green trees and multiple stages of post-insect attack tree mortality, including dead trees with red needles (“red-attack”), dead trees without needles (“gray-attack”), and non-forest. The 30-cm resolution image facilitated delineation of trees located in the field, which were used in image classification. We employed a maximum likelihood classifier using the green band, Red-Green Index (RGI), and Normalized Difference Vegetation Index (NDVI). Pixel-level classification accuracies using this imagery were good (overall accuracy of 87%, kappa = 0.84), although misclassification occurred between a) sunlit crowns of live (green) trees and herbaceous vegetation, and b) sunlit crowns of gray- and red-attack trees and bare soil. We explored the capability of coarser resolution imagery, aggregated from the 30-cm resolution to 1.2, 2.4, and 4.2 m, to improve classification accuracy. We found the highest accuracy at the 2.4-m resolution, where reduction in omission and commission errors and increases in overall accuracy (90%) and kappa (0.88) were achieved, and visual inspection indicated improved mapping. Pixels at this resolution included more shadow in forested regions than pixels in finer resolution imagery, thereby reducing forest canopy reflectance and allowing improved separation between forest and non-forest classes, yet were fine enough to resolve individual tree crowns better than the 4.2-m imagery. Our results illustrate that a classification of an image with a spatial resolution similar to the area of a tree crown outperforms that of finer and coarser resolution imagery for mapping tree mortality and non-forest classes. We also demonstrate that multispectral imagery can be used to separate multiple postoutbreak attack stages (i.e., red-attack and gray-attack) from other classes in the image.     

Integrating remotely sensed and ancillary data sources to characterize a mountain pine beetle infestation

The ongoing mountain pine beetle (Dendroctonus ponderosae Hopkins) outbreak in British Columbia, Canada, has reached epidemic proportions, with the beetle expanding into geographic areas outside its known biological range. In this study, estimates of red attack damage were derived from a logistic regression model using multi-date Landsat imagery, and ancillary information including terrain attributes and solar radiation. The model estimates were found to be approximately 70% accurate using an independent set of beetle survey data as validation. This probability surface of red attack damage, along with forest inventory and terrain attributes, were used as inputs to decision tree analyses, in order to identify which forest attributes were associated with stands that had a greater likelihood of mountain pine beetle red attack damage. Three distinct decision tree models were developed, with each having a different set of input variables. The results of the analyses indicated that site index (an indicator of the quality of a forest site) and slope were the principal discriminators of the current mountain pine beetle attack, followed by basal area of pine dominated stands, and to a lesser extent, crown closure and stem density. The results suggest that indicators of site quality, particularly site index, could be a complementary addition to existing stand susceptibility rating models.     

Assessing canopy mortality during a mountain pine beetle outbreak using GeoEye-1 high spatial resolution satellite data

Continuing, severe outbreaks of mountain pine beetle (Dendroctonus ponderosae) across western North America have resulted in widespread mortality of lodgepole pine (Pinus contorta). Multiple studies have used high spatial resolution satellite data to map areas of beetle kill; these studies have largely focused on mapping red canopy cover associated with recent tree mortality and have not examined mapping gray canopy cover that occurs after red needles have dropped. The work presented here examines the use of newly available GeoEye-1 data for mapping both red and gray canopy area in southeastern Wyoming lodgepole pine forest. A 0.5 m spatial resolution, pan-sharpened GeoEye-1 image was used to classify areas of green, red, and gray canopy cover. Reference data were collected at twelve 500 m² field plots. Shadow-normalized green, red, and gray canopy area from classified GeoEye-1 data closely agreed with field-estimated green, red, and gray canopy area. Mean absolute error in canopy cover for the twelve sample plots was 8.3% for the green class, 5.4% for the red class, and 7.2% for the gray class. When all twelve plots were aggregated, remotely sensed estimates of green, red, and gray cover were within 1.7% of the field-estimated cover. Our results demonstrate that high spatial resolution spaceborne multispectral data are a promising tool for mapping canopy mortality caused by mountain pine beetle outbreaks.     

Multi-temporal analysis of high spatial resolution imagery for disturbance monitoring

Mountain pine beetle red attack damage has been successfully detected and mapped using single-date high spatial resolution (< 4 m) satellite multi-spectral data. Forest managers; however, need to monitor locations for changes in beetle populations over time. Specifically, counts of individual trees attacked in successive years provide an indication of beetle population growth and dynamics. Surveys are typically used to estimate the ratio of green (current) attack trees to red (previous) attack trees or G:R. In this study, we estimate average stand-level G:R using a time-series of QuickBird multi-spectral and panchromatic satellite data, combined with field data for three forested stands near Merritt, British Columbia, Canada. Using a ratio of QuickBird red to green wavelengths (Red-Green Index or RGI), the change in RGI (ΔRGI) in successive image pairs is used to estimate red attack damage in 2004, 2005, and 2006, with true positive accuracies ranging from 89 to 93%. To overcome issues associated with differing viewing geometry and illumination angles that impair tracking of individual trees through time, segments are generated from the QuickBird multi-spectral data to identify small groups of trees. These segments then serve as the vehicle for monitoring changes in red attack damage over time. A local maxima filter is applied to the panchromatic data to estimate stem counts, thereby allowing an indication of the total stand population at risk of attack. By combining the red attack damage estimates with the local maxima stem counts, predictions are made of the number of attacked trees in a given year. Backcasting the current year's red attack damaged trees as the previous year's green attack facilitates the estimation of an average stand G:R. In this study area, these retrospective G:R values closely match those generated from field surveys. The results of this study indicate that a monitoring program using a time series of high spatial resolution remotely sensed data (multi-spectral and panchromatic) over select sample locations, could be used to estimate G:R over large areas, facilitating landscape level management strategies and/or providing a mechanism for assessing the efficacy of previously implemented strategies.     

Prediction and assessment of bark beetle-induced mortality of lodgepole pine using estimates of stand vigor derived from remotely sensed data

The current outbreak of mountain pine beetle (Dendroctonus ponderosae Hopkins) in British Columbia (BC), Canada, has led forest managers to consider thinning as a means of decreasing residual tree susceptibility to attack and subsequent mortality. Previous research indicates that susceptibility to mountain pine beetle is a function of a tree's physiological vigor and the intensity of attack. Trees able to produce ≥ 80 g (g) of wood per m² of projected leaf area annually are highly resistant, because they are able to shift resource allocation locally from wood to resin production to isolate blue-stain fungi introduced by attacking beetles. Typically, the leaf area of susceptible stands must be reduced by two-thirds to permit most residual trees to increase their vigor to a safe level. We evaluate whether Landsat Thematic Mapper (TM) imagery (30 × 30 m) provides a means to assess the maximum leaf area index (LAI) of unthinned stands and the extent that thinning reduces LAI. The extent that residual trees in thinned stands may have increased their resistance to attack from mountain pine beetle is predicted from a non-linear relationship between % maximum LAI and mean tree vigor.We investigated the merits of this approach in the vicinity of Parson, British Columbia using four stands of lodgepole pine (Pinus contorta Dougl.), two of which were heavily thinned (stands were spaced to 4 and 5 m, approximately 70% reduction in stand density). An analysis of archived Landsat TM imagery indicated that prior to thinning in 1993, all four stands had full canopy, which, for mature stands, would translate to mean tree vigor between 40 and 70 g of annual wood production per m² of foliage. By 1995, based on estimated changes in LAI derived from a second data of Landsat TM imagery, stand vigor in the unthinned stands had not changed; however, in the thinned stands, a nearly two third reduction in LAI resulted in a predicted increase in vigor to between 100 and 160 g wood m^− 2 of leaf area. A subsequent assessment in 2001 indicated that stand vigor remained higher in the thinned stands relative to the control stands. Following an infestation of mountain pine beetle in the study area in 2002, mortality data indicated that the thinned stands experienced no mortality relative to the unthinned stands which experienced 5.5% mortality in the initial years of the attack. In the larger area surrounding the study site, a general relationship was found between predicted stand vigor and mountain pine beetle-induced mortality as estimated from aerial overview survey data (r² = 0.43, p < 0.01).     

Detection of mountain pine beetle-killed ponderosa pine in a heterogeneous landscape using high-resolution aerial imagery

Previous studies have used remote-sensing images to map tree mortality caused by mountain pine beetle (Dendroctonus ponderosae; MPB) in relatively homogeneous lodgepole pine (Pinus contorta) forests; however, classification methods have not been tested for the patchy landscape of ponderosa pine-dominated (Pinus ponderosae) montane forests characterized by highly variable tree density. This study explores two supervised classification methods to identify MPB-caused mortality (red attack) in heterogeneous montane forests of the Colorado Front Range using 1 m-resolution 2011 imagery of the National Agriculture Imagery Program (NAIP): maximum likelihood using the red, green, and blue bands, and the red-green index (RGI), and a thresholding technique using the RGI. Two variations of the RGI threshold method were also explored: the addition of a green-band threshold and the incorporation of a focal analysis. Evaluation pixels were used to assess the accuracy of the classification methods. The maximum likelihood (97 Percentage Correctly Classified (PCC); 11% error of commission for red attack) and RGI threshold (85 PCC; 46% error of commission for red attack) classification methods overestimated the red attack. The RGI and green band threshold classification reduced the error of commission (5%) and had high overall accuracy (97 PCC). In a comparison of classification methods across tree-density sites, we found the maximum likelihood classification had a very high accuracy in the high-density site (95 PCC), but substantially lower accuracy in the low-density site (85 PCC) due to the presence of more visible cover types. The RGI threshold classification with the green band constraint produced more consistent PCCs across tree densities: high (93.7 PCC), moderate (95.2 PCC), and low (92.0 PCC). Our results indicate forest structure may affect the classification accuracy and should be considered when selecting a classification method for a landscape.     

Continuous wavelet analysis for the detection of green attack damage due to mountain pine beetle infestation

Mountain pine beetle (Dendroctonus ponderosae Hopkins) is the most destructive insect infesting mature pine forests in North America and has devastated millions of hectares of forest in western Canada. Past studies have demonstrated the use of multispectral imagery for remote identification and mapping of visible or red attack damage in forests. This study aims to detect pre-visual or green attack damage in lodgepole pine needles by means of hyperspectral measurements, particularly via continuous wavelet analysis. Field measurements of lodgepole pine stands were conducted at two sites located northwest of Edmonton, Alberta, Canada. In June and August of 2007, reflectance spectra (350-2500 nm) were collected for 16 pairs of trees. Each of the 16 tree pairs included one control tree (healthy), and one stressed tree (girdled to simulate the effects of beetle infestation). In addition, during the period of June through October 2008, spectra were collected from 15 pairs of control- and beetle-infested trees. Spectra derived from these 31 tree pairs were subjected to a continuous wavelet transform, generating a scalogram that compiles the wavelet power as a function of wavelength location and scale of decomposition. Linear relationships were then explored between the wavelet scalograms and chemical properties or class labels (control and non-control) of the sample populations in order to isolate the most useful distinguishing spectral features that related to infested or girdled trees vs. control trees.A deficit in water content is observed in infested trees while an additional deficit in chlorophyll content is seen for girdled trees. The measurable water deficit of infested and girdled tree samples was detectable from the wavelet analysis of the reflectance spectra providing a novel method for the detection of green attack. The spectral features distinguishing control and infested trees are predominantly located between 950 and 1390 nm from scales 1 to 8. Of those, five features between 1318 to 1322 nm at scale 7 are consistently found in the July and August 2008 datasets. These features are located at longer wavelengths than those investigated in previous studies (below 1100 nm) and provide new insights into the potential remote detection of green attack. Spectral features that distinguish control and girdled trees were mostly observed between 1550 and 2370 nm from scales 1 to 5. The differing response of girdled and infested trees appears to indicate that the girdling process does not provide a perfect simulation of the effects caused by beetle infestation.It remains to be determined if the location of the 1318-1322 nm features, near the edge of a strong atmospheric water absorption band, will be sufficiently separable for use in airborne detection of green attack. A plot comparing needle water content and wavelet power at 1320 nm reveals considerable overlap between data derived from both infested and control samples, though the groups are statistically separable. This obstacle may preclude a high accuracy separation of healthy and infected single individuals, but establishing threshold identification levels may provide an economical, efficient and expeditious method for discriminating between healthy and infested tree populations.     

A Landsat time series approach to characterize bark beetle and defoliator impacts on tree mortality and surface fuels in conifer forests

Insects are important forest disturbance agents, and mapping their effects on tree mortality and surface fuels represents a critical research challenge. Although various remote sensing approaches have been developed to monitor insect impacts, most studies have focused on single insect agents or single locations and have not related observed changes to ground-based measurements. This study presents a remote sensing framework to (1) characterize spectral trajectories associated with insect activity of varying duration and severity and (2) relate those trajectories to ground-based measurements of tree mortality and surface fuels in the Cascade Range, Oregon, USA. We leverage a Landsat time series change detection algorithm (LandTrendr), annual forest health aerial detection surveys (ADS), and field measurements to investigate two study landscapes broadly applicable to conifer forests and dominant insect agents of western North America. We distributed 38 plots across multiple forest types (ranging from mesic mixed-conifer to xeric lodgepole pine) and insect agents (defoliator [western spruce budworm] and bark beetle [mountain pine beetle]). Insect effects were evident in the Landsat time series as combinations of both short- and long-duration changes in the Normalized Burn Ratio spectral index. Western spruce budworm trajectories appeared to show a consistent temporal evolution of long-duration spectral decline (loss of vegetation) followed by recovery, whereas mountain pine beetle plots exhibited both short- and long-duration spectral declines and variable recovery rates. Although temporally variable, insect-affected stands generally conformed to four spectral trajectories: short-duration decline then recovery, short- then long-duration decline, long-duration decline, long-duration decline then recovery. When comparing remote sensing data with field measurements of insect impacts, we found that spectral changes were related to cover-based estimates (tree basal area mortality [R²_adj = 0.40, F_1,34 = 24.76, P < 0.0001] and down coarse woody detritus [R²_adj = 0.29, F_1,32 = 14.72, P = 0.0006]). In contrast, ADS changes were related to count-based estimates (e.g., ADS mortality from mountain pine beetle positively correlated with ground-based counts [R²_adj = 0.37, F_1,22 = 14.71, P = 0.0009]). Fine woody detritus and forest floor depth were not well correlated with Landsat- or aerial survey-based change metrics. By characterizing several distinct temporal manifestations of insect activity in conifer forests, this study demonstrates the utility of insect mapping methods that capture a wide range of spectral trajectories. This study also confirms the key role that satellite imagery can play in understanding the interactions among insects, fuels, and wildfire.     

Segment-constrained regression tree estimation of forest stand height from very high spatial resolution panchromatic imagery over a boreal environment

Mean stand height is an important parameter for forest volume and biomass estimation in support of monitoring and management activities. Information on mean stand height is typically obtained through the manual interpretation of aerial photography, often supplemented by the collection of field calibration data. In remote areas where forest management practices may not be spatially exhaustive or where it is difficult to acquire aerial photography, alternate approaches for estimating stand height are required. One approach is to use very high spatial resolution (VHSR) satellite imagery (pixels sided less than 1 m) as a surrogate for air photos. In this research we demonstrate an approach for modelling mean stand height at four sites in the Yukon Territory, Canada, from QuickBird panchromatic imagery. An object-based approach was used to generate homogenous segments from the imagery (analogous to manually delineated forest stands) and an algorithm was used to automatically delineate individual tree crowns within the segments. A regression tree was used to predict mean stand height from stand-level metrics generated from the image grey-levels and within-stand objects relating individual tree crown characteristics. Heights were manually interpreted from the QuickBird imagery and divided into separate sets of calibration and validation data. The effects of calibration data set size and the input metrics used on the regression tree results were also assessed. The approach resulted in a model with a significant R² of 0.53 and an RMSE of 2.84 m. In addition, 84.6% of the stand height estimates were within the acceptable error for photo interpreted heights, as specified by the forest inventory standards of British Columbia. Furthermore, residual errors from the model were smallest for the stands that had larger mean heights (i.e., > 20 m), which aids in reducing error in subsequent estimates of biomass or volume (since stands with larger trees contribute more to overall estimates of volume or biomass). Estimated and manually interpreted heights were reclassified into 5-metre height classes (a schema frequently used for forest analysis and modelling applications) and compared; classes corresponded in 54% of stands assessed, and all stands had an estimated height class that was within ± 1 class of their actual class. This study demonstrates the capacity of VHSR panchromatic imagery (in this case QuickBird) for generating useful estimates of mean stand heights in unmonitored, remote, or inaccessible forest areas.     

Detection of foliage conditions and disturbance from multi-angular high spectral resolution remote sensing

Disturbance of forest ecosystems, an important component of the terrestrial carbon cycle, has become a focus of research over recent years, as global warming is about to increase the frequency and severity of natural disturbance events. Remote sensing offers unique opportunities for detection of forest disturbance at multiple scales; however, spatially and temporally continuous mapping of non-stand replacing disturbance remains challenging. First, most high spatial resolution satellite sensors have relatively broad spectral ranges with bandwidths unsuitable for detection of subtle, stress induced, features in canopy reflectance. Second, directional and background reflectance effects, induced by the interactions between the sun-sensor geometry and the observed canopy surface, make up-scaling of empirically derived relationships between changes in spectral reflectance and vegetation conditions difficult. Using an automated tower based spectroradiometer, we analyse the interactions between canopy level reflectance and different stages of disturbance occurring in a mountain pine beetle infested lodgepole pine stand in northern interior British Columbia, Canada, during the 2007 growing season. Directional reflectance effects were modelled using a bidirectional reflectance distribution function (BRDF) acquired from high frequency multi-angular spectral observations. Key wavebands for observing changes in directionally corrected canopy spectra were identified using discriminant analysis and highly significant correlations between canopy reflectance and field measured disturbance levels were found for several broad and narrow waveband vegetation indices (for instance, r²_NDVI = 0.90; r²_CHL3 = 0.85; p < 0.05). Results indicate that multi-angular observations are useful for extraction of disturbance related changes in canopy reflectance, in particular the temporally and spectrally dense data detected changes in chlorophyll content well. This study will help guide and inform future efforts to map forest health conditions at landscape and over increasingly coarse scales.     

Forest carbon densities and uncertainties from Lidar, QuickBird, and field measurements in California

Greenhouse gas inventories and emissions reduction programs require robust methods to quantify carbon sequestration in forests. We compare forest carbon estimates from Light Detection and Ranging (Lidar) data and QuickBird high-resolution satellite images, calibrated and validated by field measurements of individual trees. We conducted the tests at two sites in California: (1) 59 km² of secondary and old-growth coast redwood (Sequoia sempervirens) forest (Garcia-Mailliard area) and (2) 58 km² of old-growth Sierra Nevada forest (North Yuba area). Regression of aboveground live tree carbon density, calculated from field measurements, against Lidar height metrics and against QuickBird-derived tree crown diameter generated equations of carbon density as a function of the remote sensing parameters. Employing Monte Carlo methods, we quantified uncertainties of forest carbon estimates from uncertainties in field measurements, remote sensing accuracy, biomass regression equations, and spatial autocorrelation. Validation of QuickBird crown diameters against field measurements of the same trees showed significant correlation (r = 0.82, P < 0.05). Comparison of stand-level Lidar height metrics with field-derived Lorey's mean height showed significant correlation (Garcia-Mailliard r = 0.94, P < 0.0001; North Yuba R = 0.89, P < 0.0001). Field measurements of five aboveground carbon pools (live trees, dead trees, shrubs, coarse woody debris, and litter) yielded aboveground carbon densities (mean ± standard error without Monte Carlo) as high as 320 ± 35 Mg ha^− 1 (old-growth coast redwood) and 510 ± 120 Mg ha^− 1 (red fir [Abies magnifica] forest), as great or greater than tropical rainforest. Lidar and QuickBird detected aboveground carbon in live trees, 70-97% of the total. Large sample sizes in the Monte Carlo analyses of remote sensing data generated low estimates of uncertainty. Lidar showed lower uncertainty and higher accuracy than QuickBird, due to high correlation of biomass to height and undercounting of trees by the crown detection algorithm. Lidar achieved uncertainties of < 1%, providing estimates of aboveground live tree carbon density (mean ± 95% confidence interval with Monte Carlo) of 82 ± 0.7 Mg ha^− 1 in Garcia-Mailliard and 140 ± 0.9 Mg ha^− 1 in North Yuba. The method that we tested, combining field measurements, Lidar, and Monte Carlo, can produce robust wall-to-wall spatial data on forest carbon.     

Mapping of riparian zone attributes using discrete return LiDAR, QuickBird and SPOT-5 imagery: Assessing accuracy and costs

Riparian zones in Australia are exposed to increasing pressures because of disturbance from agricultural and urban expansion, weed invasion, and overgrazing. Accurate and cost-effective mapping of riparian environments is important for assessing riparian zone functions associated with water quality, biodiversity, and wildlife habitats. The objective of this research was to compare the accuracy and costs of mapping riparian zone attributes from image data acquired by three different sensor types, i.e. Light Detection and Ranging (LiDAR) (0.5-2.4 m pixels), and multi-spectral QuickBird (2.4 m pixels) and SPOT-5 (10 m pixels). These attributes included streambed width, riparian zone width, plant projective cover, longitudinal continuity, vegetation overhang, and bank stability. The riparian zone attributes were mapped for a study area along Mimosa Creek in the Fitzroy Catchment, Central Queensland, Australia. Object-based image and regression analyses were used for mapping the riparian zone attributes. The validation of the LiDAR, QuickBird, and SPOT-5 derived maps of streambed width (R = 0.99, 0.71, and 0.44 respectively) and riparian zone width (R = 0.91, 0.87, and 0.74 respectively) against field derived measurements produced the highest accuracies for the LiDAR data and the lowest using the SPOT-5 image data. Cross-validation estimates of misclassification produced a root mean square error of 1.06, 1.35 and 1.51 from an ordinal scale from 0 to 4 of the bank stability score for the LiDAR, QuickBird and SPOT-5 image data, respectively. The validation and empirical modelling showed high correlations for all datasets for mapping plant projective cover (R > 0.93). The SPOT-5 image data were unsuitable for assessment of riparian zone attributes at the spatial scale of Mimosa Creek and associated riparian zones. Cost estimates of image and field data acquisition and processing of the LiDAR, QuickBird, and SPOT-5 image data showed that discrete return LiDAR can be used for costs lower than those for QuickBird image data over large spatial extents (e.g. 26,000 km of streams). With the higher level of vegetation structural and landform information, mapping accuracies, geometric precision, and lower overall costs at large spatial extents, LiDAR data are a feasible means for assessment of riparian zone attributes.     

Use of high-resolution satellite imagery in an integrated model to predict the distribution of shade coffee tree hybrid zones

In New Caledonia (21°S, 165°E), shade-grown coffee plantations were abandoned for economic reasons in the middle of the 20th century. Coffee species (Coffea arabica, C. canephora and C. liberica) were introduced from Africa in the late 19th century, they survived in the wild and spontaneously cross-hybridized. Coffee species were originally planted in native forest in association with leguminous trees (mostly introduced species) to improve their growth. Thus the canopy cover over rustic shade coffee plantations is heterogeneous with a majority of large crowns, attributed to leguminous trees. The aim of this study was to identify suitable areas for coffee inter-specific hybridization in New Caledonia using field based environmental parameters and remotely sensed predictors. Due to the complex structure of tropical vegetation, remote sensing imagery needs to be spatially accurate and to have the appropriate bands for monitoring vegetation cover. Quickbird panchromatic (black and white) imagery at 0.6 to 0.7 m spatial resolutions and multispectral imagery at 2.4 m spatial resolution were pansharpened and used for this study. The two most suitable remotely sensed indicators, canopy heterogeneity and tree crown size, were acquired by the sequential use of tree crown detection (neural network), image processing (such as textural analysis) and classification. All models were supervised and trained on learning data determined by human expertise. The final model has two remotely sensed indicators and three physical parameters based on the Digital Elevation Model: elevation, slope and water flow accumulation. Using these five predictive variables as inputs, two modelling methods, a decision tree and a neural network, were implemented. The decision tree, which showed 96.9% accuracy on the test set, revealed the involvement of ecological parameters in the hybridization of Coffea species. We showed that hybrid zones could be characterized by combinations of modalities, underlining the complexity of the environment concerned. For instance, forest heterogeneity and large crown size, steep slopes (> 53.5%) and elevation between 194 and 429 m asl, are favourable factors for Coffea inter-specific hybridization. The application of the neural network on the whole area gave a predictive map that distinguished the most suitable areas by means of a nonlinear continuous indicator. The map provides a confidence level for each area. The most favourable areas were geographically localized, providing a clue for the detection and conservation of favourable areas for Coffea species neo-diversity.     

Sensitivity of the thematic mapper enhanced wetness difference index to detect mountain pine beetle red-attack damage

Red-attack damage caused by mountain pine beetle (Dentroctonus ponderosa Hopkins) infestation in stands of lodgepole pine (Pinus contorta) in the Prince George Forest Region of British Columbia was examined using multitemporal Landsat-7 ETM+ imagery acquired in 1999, 2000, and 2001. The image data were geometrically and atmospherically corrected, and processed using the Tasseled Cap Transformation (TCT) to obtain wetness indices. The final steps included pixel subtraction, enhancement, and thresholding of the wetness index differences. The resulting enhanced wetness difference index (EWDI) was used to interpret spectral patterns in stands with confirmed (through aerial survey) red-attack damage in 2001, and these EWDI patterns were compared to the patterns of reflectance in normal-colour composites. We stratified the aerial survey dataset into two levels and used the EWDI to discriminate classes of 10-29 red-attack trees and 30-50 red-attack trees, and a sample of healthy forest collected from inventory data. Classification accuracy of red-attack damage based on the EWDI ranged from 67% to 78% correct.     

Using hyperspectral remote sensing to detect and quantify southeastern pine senescence effects in red-cockaded woodpecker (Picoides borealis) habitat

Conservation of threatened and endangered species requires maintenance of critical habitat. The red-cockaded woodpecker Picoides borealis (RCW) is a threatened bird species, endemic to the mixed conifer forests of the southeastern United States. RCW nests and forages preferentially in mature longleaf pine Pinus palustris, but also uses mature loblolly pine Pinus taeda and shortleaf pine Pinus echinata forests. In the last century, the extent of mature pine forests has been greatly reduced by logging. The RCW, in contrast to other woodpeckers, excavates nest cavities in living trees and senescence symptoms (year round leaf chlorosis and leaf mortality) have been observed in mature pine stands across the southeast. Widespread mortality of the mature pine forests would threaten the long-term survival of the RCW. We used airborne hyperspectral data across a portion of Ft. Benning Military Installation, Georgia, U.S.A., to determine if senescent trees can be identified and mapped and assess the likely persistence of mature pines in the RCW habitat. Univariate analysis of variance showed good separation between asymptomatic, senesced and dead physiological conditions with asymptomatic trees having significantly higher reflectance for all bands in the wavelength range between 0.719 and 1.1676 µm, senescent trees having significantly lower reflectance for bands in the range between 1.1927 and 1.3122 µm, and dead trees having significantly higher reflectance for bands in the range between 1.8151 and 1.9471 µm. Classification and Regression Tree (CART) models achieved correct classification rates and kappa statistics above 70%. CART models selected information from wavelength regions similar to those identified from the ANOVA, which likely explains their performance. Our aggregated CART model of tree senescence estimated that 141.4 ha (3%) of the study area is affected. RCW nests occurred in areas with significantly higher tree cover, and trees within foraging and home ranges did not have significantly more senescence than areas without RCW. These results indicate that although tree senescence is widespread, mortality is yet to significantly affect RCW habitat. Results and analysis of critical habitat similar to those exemplified in this study can extend our knowledge about the stressors of these important and imperiled components of biodiversity.     

Mapping whitebark pine mortality caused by a mountain pine beetle outbreak with high spatial resolution satellite imagery

Insect outbreaks cause significant tree mortality across western North America, including in high-elevation whitebark pine forests. These forests are under several threats, which include attack by insects and white pine blister rust, as well as conversion to other tree species as a result of fire suppression. Mapping tree mortality is critical to determining the status of whitebark pine as a species. Satellite remote sensing builds upon existing aerial surveys by using objective, repeatable methods that can result in high spatial resolution monitoring. Past studies concentrated on level terrain and only forest vegetation type. The objective of this study was to develop a means of classifying whitebark pine mortality caused by a mountain pine beetle infestation in rugged, remote terrain using high spatial resolution satellite imagery. We overcame three challenges of mapping mortality in this mountainous region: (1) separating non-vegetated cover types, green and brown herbaceous cover, green (live) tree cover, and red-attack (dead) tree cover; (2) variations in illumination as a result of variations in slope and aspect related to the mountainous terrain of the study site; and (3) the difficulty of georegistering the imagery for use in comparing field measurements. Quickbird multi-spectral imagery (2.4 m spatial resolution) was used, together with a maximum likelihood classification method, to classify vegetation cover types over a 6400 ha area. To train the classifier, we selected pixels in each cover class from the imagery guided by our knowledge of the study site. Variables used in the maximum likelihood classifier included the ratio of red reflectance to green reflectance as well as green reflectance. These variables were stratified by solar incidence angle to account for illumination variability. We evaluated the results of the classified image using a reserved set of image-derived class members and field measurements of live and dead trees. Classification results yielded high overall accuracy (86% and 91% using image-derived class members and field measurements respectively) and kappa statistics (0.82 and 0.82) and low commission (0.9% and 1.5%) and omission (6.5% and 15.9%) errors for the red-attack tree class. Across the scene, 700 ha or 31% of the forest was identified as in the red-attack stage. Severity (percent mortality by canopy cover) varied from nearly 100% for some areas to regions with little mortality. These results suggest that high spatial resolution satellite imagery can provide valuable information for mapping and monitoring tree mortality even in rugged, mountainous terrain.     

Estimating aboveground carbon in a catchment of the Siberian forest tundra: Combining satellite imagery and field inventory

This study was part of an interdisciplinary research project on soil carbon and phytomass dynamics of boreal and arctic permafrost landscapes. The 45 ha study area was a catchment located in the forest tundra in northern Siberia, approximately 100 km north of the Arctic Circle.The objective of this study was to estimate aboveground carbon (AGC) and assess and model its spatial variability. We combined multi-spectral high resolution remote sensing imagery and sample based field inventory data by means of the k-nearest neighbor (k-NN) technique and linear regression.Field data was collected by stratified systematic sampling in August 2006 with a total sample size of n = 31 circular nested sample plots of 154 m² for trees and shrubs and 1 m² for ground vegetation. Destructive biomass samples were taken on a sub-sample for fresh weight and moisture content. Species-specific allometric biomass models were constructed to predict dry biomass from diameter at breast height (dbh) for trees and from elliptic projection areas for shrubs.Quickbird data (standard imagery product), acquired shortly before the field campaign and archived ASTER data (Level-1B product) of 2001 were geo-referenced, converted to calibrated radiances at sensor and used as carrier data. Spectral information of the pixels which were located in the inventory plots were extracted and analyzed as reference set. Stepwise multiple linear regression was applied to identify suitable predictors from the set of variables of the original satellite bands, vegetation indices and texture metrics. To produce thematic carbon maps, carbon values were predicted for all pixels of the investigated satellite scenes. For this prediction, we compared the kNN distance-weighted classifier and multiple linear regression with respect to their predictions.The estimated mean value of aboveground carbon from stratified sampling in the field is 15.3 t/ha (standard error SE = 1.50 t/ha, SE% = 9.8%). Zonal prediction from the k-NN method for the Quickbird image as carrier is 14.7 t/ha with a root mean square error RMSE = 6.42 t/ha, RMSE_r = 44%) resulting from leave-one-out cross-validation. The k-NN-approach allows mapping and analysis of the spatial variability of AGC. The results show high spatial variability with AGC predictions ranging from 4.3 t/ha to 28.8 t/ha, reflecting the highly heterogeneous conditions in those permafrost-influenced landscapes. The means and totals of linear regression and k-NN predictions revealed only small differences but some regional distinctions were recognized in the maps.     

Leaf chlorophyll content retrieval from airborne hyperspectral remote sensing imagery

Hyperspectral remote sensing has great potential for accurate retrieval of forest biochemical parameters. In this paper, a hyperspectral remote sensing algorithm is developed to retrieve total leaf chlorophyll content for both open spruce and closed forests, and tested for open forest canopies. Ten black spruce (Picea mariana (Mill.)) stands near Sudbury, Ontario, Canada, were selected as study sites, where extensive field and laboratory measurements were carried out to collect forest structural parameters, needle and forest background optical properties, and needle biophysical parameters and biochemical contents chlorophyll a and b. Airborne hyperspectral remote sensing imagery was acquired, within one week of ground measurements, by the Compact Airborne Spectrographic Imager (CASI) in a hyperspectral mode, with 72 bands and half bandwidth 4.25-4.36 nm in the visible and near-infrared region and a 2 m spatial resolution. The geometrical-optical model 4-Scale and the modified leaf optical model PROSPECT were combined to estimate leaf chlorophyll content from the CASI imagery. Forest canopy reflectance was first estimated with the measured leaf reflectance and transmittance spectra, forest background reflectance, CASI acquisition parameters, and a set of stand parameters as inputs to 4-Scale. The estimated canopy reflectance agrees well with the CASI measured reflectance in the chlorophyll absorption sensitive regions, with discrepancies of 0.06%-1.07% and 0.36%-1.63%, respectively, in the average reflectances of the red and red-edge region. A look-up-table approach was developed to provide the probabilities of viewing the sunlit foliage and background, and to determine a spectral multiple scattering factor as functions of leaf area index, view zenith angle, and solar zenith angle. With the look-up tables, the 4-Scale model was inverted to estimate leaf reflectance spectra from hyperspectral remote sensing imagery. Good agreements were obtained between the inverted and measured leaf reflectance spectra across the visible and near-infrared region, with R² = 0.89 to R² = 0.97 and discrepancies of 0.02%-3.63% and 0.24%-7.88% in the average red and red-edge reflectances, respectively. Leaf chlorophyll content was estimated from the retrieved leaf reflectance spectra using the modified PROSPECT inversion model, with R² = 0.47, RMSE = 4.34 μg/cm², and jackknifed RMSE of 5.69 μg/cm² for needle chlorophyll content ranging from 24.9 μg/cm² to 37.6 μg/cm². The estimates were also assessed at leaf and canopy scales using chlorophyll spectral indices TCARI/OSAVI and MTCI. An empirical relationship of simple ratio derived from the CASI imagery to the ground-measured leaf area index was developed (R² = 0.88) to map leaf area index. Canopy chlorophyll content per unit ground surface area was then estimated, based on the spatial distributions of leaf chlorophyll content per unit leaf area and the leaf area index.     

Estimation of insect infestation dynamics using a temporal sequence of Landsat data

The current outbreak of mountain pine beetle (Dendroctonus ponderosae) in western Canada has been increasing over the past decade and is currently estimated to be impacting 9.2 million hectares, with varying levels of severity. Large area insect monitoring is typically undertaken using manual aerial overview sketch mapping, whereby an interpreter depicts areas of homogenous insect attack conditions onto 1:250,000 or 1:100,000 scale maps. These surveys provide valuable strategic data for management at the provincial scale. The coarse spatial and attribute resolution of these data however, make them inappropriate for fine-scale monitoring and operational planning. For instance, it is not possible to estimate the initial timing of attack and year of stand death. In this study, we utilise eight Landsat scenes collected over a 14 year period in north-central British Columbia, Canada, where the infestation has gradually developed both spatially and temporally. After pre-processing and normalising the eight scenes using a relative normalisation procedure, decision tree analysis was applied to classify spectral trajectories of the Normalised Difference Moisture Index (NDMI). From the classified temporal sequence of images, key parameters were extracted including the presence of beetle disturbance and timing of stand decline. The accuracy of discriminating beetle attack from healthy forest stands was assessed both spatially and temporally using three years of aerial survey data (1996, 2003, and 2004) with results indicating overall classification accuracies varying between 71 and 86%. As expected, the earliest and least severe attack year (1996), recorded the lowest overall accuracy. The relationship between the timing of stand attack (i.e. moderate to severe beetle infestation) and NDMI (initial year of detected disturbance) was also explored. The results suggest that there is potential for deriving regional estimates of the year of stand death using Landsat data and decision tree analysis however, a higher temporal frequency of images is required to quantify the timing of mountain pine beetle attack.     

Generation of dense time series synthetic Landsat data through data blending with MODIS using a spatial and temporal adaptive reflectance fusion model

Landsat imagery with a 30 m spatial resolution is well suited for characterizing landscape-level forest structure and dynamics. While Landsat images have advantageous spatial and spectral characteristics for describing vegetation properties, the Landsat sensor's revisit rate, or the temporal resolution of the data, is 16 days. When considering that cloud cover may impact any given acquisition, this lengthy revisit rate often results in a dearth of imagery for a desired time interval (e.g., month, growing season, or year) especially for areas at higher latitudes with shorter growing seasons. In contrast, MODIS (MODerate-resolution Imaging Spectroradiometer) has a high temporal resolution, covering the Earth up to multiple times per day, and depending on the spectral characteristics of interest, MODIS data have spatial resolutions of 250 m, 500 m, and 1000 m. By combining Landsat and MODIS data, we are able to capitalize on the spatial detail of Landsat and the temporal regularity of MODIS acquisitions. In this research, we apply and demonstrate a data fusion approach (Spatial and Temporal Adaptive Reflectance Fusion Model, STARFM) at a mainly coniferous study area in central British Columbia, Canada. Reflectance data for selected MODIS channels, all of which were resampled to 500 m, and Landsat (at 30 m) were combined to produce 18 synthetic Landsat images encompassing the 2001 growing season (May to October). We compared, on a channel-by-channel basis, the surface reflectance values (stratified by broad land cover types) of four real Landsat images with the corresponding closest date of synthetic Landsat imagery, and found no significant difference between real (observed) and synthetic (predicted) reflectance values (mean difference in reflectance: mixed forest x? = 0.086, σ = 0.088, broadleaf x? = 0.019, σ = 0.079, coniferous x? = 0.039, σ = 0.093). Similarly, a pixel based analysis shows that predicted and observed reflectance values for the four Landsat dates were closely related (mean r² = 0.76 for the NIR band; r² = 0.54 for the red band; p < 0.01). Investigating the trend in NDVI values in synthetic Landsat values over a growing season revealed that phenological patterns were well captured; however, when seasonal differences lead to a change in land cover (i.e., disturbance, snow cover), the algorithm used to generate the synthetic Landsat images was, as expected, less effective at predicting reflectance.