High-resolution image texture as a predictor of bird species richness

We tested image texture as a predictor of bird species richness in a semi-arid landscape of New Mexico. Bird species richness was summarized from 10-min point counts conducted at 12 points within 42 plots (108 ha each) from 1996 to 1998. We calculated 14 first- and second-order texture measures in eight different window sizes on a set of digital orthophotos acquired in 1996. For each of the 42 plots, we summarized mean and standard deviation of each texture value within multiple window sizes. The relationship between image texture and average bird species richness was assessed using linear regression models. Single image texture measures such as the standard deviation described up to 57% of the variability in species richness. Coupling multiple measures of texture or coupling elevation with a single texture measure described up to 63% of the variability in bird species richness. Models incorporating two measures of texture and coarse habitat type described 76% of the variability in bird species richness. These results show that image texture analysis is a very promising tool for characterizing habitat structure and predicting patterns of species richness in semi-arid ecosystems. This method has several advantages over methods that rely on classified imagery, including cost-effectiveness, incorporation of within-habitat vegetation variability, and elimination of errors associated with boundary delineation.     

Modeling forest songbird species richness using LiDAR-derived measures of forest structure

Conservation of biodiversity requires information at many spatial scales in order to detect and preserve habitat for many species, often simultaneously. Vegetation structure information is particularly important for avian habitat models and has largely been unavailable for large areas at the desired resolution. Airborne LiDAR, with its combination of relatively broad coverage and fine resolution provides existing new opportunities to map vegetation structure and hence avian habitat. Our goal was to model the richness of forest songbirds using forest structure information obtained from LiDAR data. In deciduous forests of southern Wisconsin, USA, we used discrete-return airborne LiDAR to derive forest structure metrics related to the height and density of vegetation returns, as well as composite variables that captured major forest structural elements. We conducted point counts to determine total forest songbird richness and the richness of foraging, nesting, and forest edge-related habitat guilds. A suite of 35 LiDAR variables were used to model bird species richness using best-subsets regression and we used hierarchical partitioning analysis to quantify the explanatory power of each variable in the multivariate models. Songbird species richness was correlated most strongly with LiDAR variables related to canopy and midstory height and midstory density (R² = 0.204, p < 0.001). Richness of species that nest in the midstory was best explained by canopy height variables (R² = 0.197, p < 0.001). Species that forage on the ground responded to mean canopy height and the height of the lower canopy (R² = 0.149, p < 0.005) while aerial foragers had higher richness where the canopy was tall and dense and the midstory more sparse (R² = 0.216, p < 0.001). Richness of edge-preferring species was greater where there were fewer vegetation returns but higher density in the understory (R² = 0.153, p < 0.005). Forest interior specialists responded positively to a tall canopy, developed midstory, and a higher proportion of vegetation returns (R² = 0.195, p < 0.001). LiDAR forest structure metrics explained between 15 and 20% of the variability in richness within deciduous forest songbird communities. This variability was associated with vertical structure alone and shows how LiDAR can provide a source of complementary predictive data that can be incorporated in models of wildlife habitat associations across broad geographical extents.     

Mapping forest patches and scattered trees from SPOT images and testing their ecological importance for woodland birds in a fragmented agricultural landscape

Scattered trees represent an important element within the agricultural matrix for birds. The aims of this study were to develop methods for mapping isolated trees from satellite imagery and to assess the importance of isolated trees for bird species richness. Field sampling of birds and plants was conducted at 120 sites in south-east Australia. We mapped tree cover from Landsat and SPOT images using a combination of spectral and segmentation based methods. Mapping of isolated trees as point objects was highly accurate (80–90%). Tree cover at spatial extents of 3–28 ha around sites explained 60% of the variability in woodland–dependent bird species richness. However, isolated trees in agricultural areas made just a small contribution to explaining the spatial variability in overall avian richness. This approach can be used for more extensive assessment of avian habitat quality from high spatial resolution images across a range of human modified landscapes.     

Using remote sensing to assess biodiversity

This review paper evaluates the potential of remote sensing for assessing species diversity, an increasingly urgent task. Existing studies of species distribution patterns using remote sensing can be essentially categorized into three types. The first involves direct mapping of individual plants or associations of single species in relatively large, spatially contiguous units. The second technique involves habitat mapping using remotely sensed data, and predictions of species distribution based on habitat requirements. Finally, establishment of direct relationships between spectral radiance values recorded from remote sensors and species distribution patterns recorded from field observations may assist in assessing species diversity. Direct mapping is applicable over smaller extents, for detailed information on the distribution of certain canopy tree species or associations. Estimations of relationships between spectral values and species distributions may be useful for the limited purpose of indicating areas with higher levels of species diversity, and can be applied over spatial extents of hundreds of square kilometres. Habitat maps appear most capable of providing information on the distributions of large numbers of species in a wider variety of habitat types. This is strongly limited by variation in species composition, and best applied over limited spatial extents of tens of square kilometres.     

Mapping the richness and composition of British breeding birds from coarse spatial resolution satellite sensor imagery

Remote sensing has great potential as a source of information on biodiversity over large areas. Past studies have generally focused on species richness, used aspatial statistical techniques and highlighted scale‐dependent results. Here, a fuller assessment of avian biodiversity, considering species richness and composition, was undertaken for breeding bird species in Great Britain from NDVI and temperature data derived from NOAA AVHRR imagery. Broad classes of bird species composition defined by an ordination analysis exhibited a high degree of separability, with classification accuracies (based on training data) of up to 77.3% observed. Although only 18.1% of the variance in species richness could be explained by a conventional aspatial regression analysis it was apparent from geographically weighted regression analyses that the relationship between species richness and the remotely sensed response was significantly non‐stationary. Relative to the standard regression, geographically weighted analyses yielded models that provided stronger relationships and highlighted the spatial dependence of the relationship. Marked spatial variation in the regression model parameters and explanatory power were evident within and between scales. The results indicated the ability to characterize aspects of biodiversity from coarse spatial resolution remote sensing data and highlight the need to accommodate for the effects of spatial non‐stationarity in the relationship.     

Mapping understory vegetation using phenological characteristics derived from remotely sensed data

Understory vegetation is an important component in forest ecosystems not only because of its contributions to forest structure, function and species composition, but also due to its essential role in supporting wildlife species and ecosystem services. Therefore, understanding the spatio-temporal dynamics of understory vegetation is essential for management and conservation. Nevertheless, detailed information on the distribution of understory vegetation across large spatial extents is usually unavailable, due to the interference of overstory canopy on the remote detection of understory vegetation. While many efforts have been made to overcome this challenge, mapping understory vegetation across large spatial extents is still limited due to a lack of generality of the developed methods and limited availability of required remotely sensed data. In this study, we used understory bamboo in Wolong Nature Reserve, China as a case study to develop and test an effective and practical remote sensing approach for mapping understory vegetation. Using phenology metrics generated from a time series of Moderate Resolution Imaging Spectroradiometer data, we characterized the phenological features of forests with understory bamboo. Using maximum entropy modeling together with these phenology metrics, we successfully mapped the spatial distribution of understory bamboo (kappa: 0.59; AUC: 0.85). In addition, by incorporating elevation information we further mapped the distribution of two individual bamboo species, Bashania faberi and Fargesia robusta (kappa: 0.68 and 0.70; AUC: 0.91 and 0.92, respectively). Due to its generality, flexibility and extensibility, this approach constitutes an improvement to the remote detection of understory vegetation, making it suitable for mapping different understory species in different geographic settings. Both biodiversity conservation and wildlife habitat management may benefit from the detailed information on understory vegetation across large areas through the applications of this approach.     

Conefor Sensinode 2.2: A software package for quantifying the importance of habitat patches for landscape connectivity

Maintaining and restoring landscape connectivity is currently a central concern in ecology and biodiversity conservation, and there is an increasing demand of user-driven tools for integrating connectivity in landscape planning. Here we describe the new Conefor Sensinode 2.2 (CS22) software, which quantifies the importance of habitat patches for maintaining or improving functional landscape connectivity and is conceived as a tool for decision-making support in landscape planning and habitat conservation. CS22 is based on graph structures, which have been suggested to possess the greatest benefit to effort ratio for conservation problems regarding landscape connectivity. CS22 includes new connectivity metrics based on the habitat availability concept, which considers a patch itself as a space where connectivity occurs, integrating in a single measure the connected habitat area existing within the patches with the area made available by the connections between different habitat patches. These new metrics have been shown to present improved properties compared to other existing metrics and are particularly suited to the identification of critical landscape elements for connectivity. CS22 is distributed together with GIS extensions that allow for directly generating the required input files from a GIS layer. CS22 and related documentation can be freely downloaded from the World Wide Web.     

Land cover heterogeneity in Great Britain as identified in Land Cover Map 2000

Land cover is an important component of landscape character. The spatial configuration and heterogeneity of land cover can influence species distributions and patterns of biodiversity. Sustainable countryside planning and land management policy in Great Britain require the mapping of land cover heterogeneity at a national scale. Here, we map four measures of land cover heterogeneity across Great Britain using Land Cover Map 2000. We calculate land cover richness, diversity, evenness and similarity within and between 1 × 1 km grid cells of the British National Grid. From this we are able to identify assemblages of land cover types that are associated with high or low landscape heterogeneity, and where they occur geographically.     

Laser remote sensing of canopy habitat heterogeneity as a predictor of bird species richness in an eastern temperate forest, USA

Habitat heterogeneity has long been recognized as a fundamental variable indicative of species diversity, in terms of both richness and abundance. Satellite remote sensing data sets can be useful for quantifying habitat heterogeneity across a range of spatial scales. Past remote sensing analyses of species diversity have largely been limited to correlative studies based on the use of vegetation indices or derived land cover maps. A relatively new form of laser remote sensing (lidar) provides another means to acquire information on habitat heterogeneity. Here we examine the efficacy of lidar metrics of canopy structural diversity as predictors of bird species richness in the temperate forests of Maryland, USA. Canopy height, topography and the vertical distribution of canopy elements were derived from lidar imagery of the Patuxent National Wildlife Refuge and compared to bird survey data collected at referenced grid locations. The canopy vertical distribution information was consistently found to be the strongest predictor of species richness, and this was predicted best when stratified into guilds dominated by forest, scrub, suburban and wetland species. Similar lidar variables were selected as primary predictors across guilds. Generalized linear and additive models, as well as binary hierarchical regression trees produced similar results. The lidar metrics were also consistently better predictors than traditional remotely sensed variables such as canopy cover, indicating that lidar provides a valuable resource for biodiversity research applications.     

Using satellite remote sensing analysis to evaluate a socio-economic and ecological model of deforestation in Rondônia,Brazil

The effectiveness of an integrated socio-economic and ecological simulation model for estimating patterns and rates of deforestation in Rondônia, Brazil is evaluated using Landsat data and landscape pattern metrics. The Percent Cleared, Contagion, and Fractal Dimension of image classifications are compared to those determined from model outputs. Results indicate that rates and spatial patterns of deforestation are similar between model outputs and Landsat image analysis. Differences in clearing patterns between the model and Landsat data are due in part to topography, localized farming obstacles and the patchiness of clearings. The effects of varying spatial resolution on the metrics is also examined.     

Mapping seagrass species, cover and biomass in shallow waters: An assessment of satellite multi-spectral and airborne hyper-spectral imaging systems in Moreton Bay (Australia)

Understanding, monitoring and modelling attributes of seagrass biodiversity, such as species composition, richness, abundance, spatial patterns, and disturbance dynamics, requires spatial information. This work assessed the accuracy of commonly available airborne hyper-spectral and satellite multi-spectral image data sets for mapping seagrass species composition, horizontal horizontal-projected foliage cover and above-ground dry-weight biomass. The work was carried out on the Eastern Banks in Moreton Bay, Australia, an area of shallow and clear coastal waters, containing a range of seagrass species, cover and biomass levels. Two types of satellite image data were used: Quickbird-2 multi-spectral and Landsat-5 Thematic Mapper multi-spectral. Airborne hyper-spectral image data were acquired from a CASI-2 sensor using a pixel size of 4.0 m. The mapping was constrained to depths shallower than 3.0 m, based on past modelling of the separability of seagrass reflectance signatures at increasing water depths. Our results demonstrated that mapping of seagrass cover, species and biomass to high accuracy levels (> 80%) was not possible across all image types. For each parameter mapped, airborne hyper-spectral data produced the highest overall accuracies (46%), followed by Quickbird-2 and then Landsat-5 Thematic Mapper. The low accuracy levels were attributed to the mapping methods and difficulties in matching locations on image and field data sets. Accurate mapping of seagrass cover, species composition and biomass, using simple approaches, requires further work using high-spatial resolution (< 5 m) and/or hyper-spectral image data. Further work is required to determine if and how the seagrass maps produced in this work are suitable for measuring attributes of seagrass biodiversity, and using these data for modelling floral and fauna biodiversity properties of seagrass environments, and for scaling-up seagrass ecosystem models.     

Predictive mapping of reef fish species richness, diversity and biomass in Zanzibar using IKONOS imagery and machine-learning techniques

During the last three decades, the large spatial coverage of remote sensing data has been used in coral reef research to map dominant substrate types, geomorphologic zones, and bathymetry. During the same period, field studies have documented statistical relationships between variables quantifying aspects of the reef habitat and its fish community. Although the results of these studies are ambiguous, some habitat variables have frequently been found to correlate with one or more aspects of the fish community. Several of these habitat variables, including depth, the structural complexity of the substrate, and live coral cover, are possible to estimate with remote sensing data. In this study, we combine a set of statistical and machine-learning models with habitat variables derived from IKONOS data to produce spatially explicit predictions of the species richness, biomass, and diversity of the fish community around two reefs in Zanzibar. In the process, we assess the ability of IKONOS imagery to estimate live coral cover, structural complexity and habitat diversity, and we explore the importance of habitat variables, at a range of spatial scales, in the predictive models using a permutation-based technique. Our findings indicate that structural complexity at a fine spatial scale (∼ 5 to 10 m) is the most important habitat variable in predictive models of fish species richness and diversity, whereas other variables such as depth, habitat diversity, and structural complexity at coarser spatial scales contribute to predictions of biomass. In addition, our results demonstrate that complex model types such as tree-based ensemble techniques provide superior predictive performance compared to the more frequently used linear models, achieving a reduction of the cross-validated root-mean-squared prediction error of 3-11%. Although aerial photographs and airborne lidar instruments have recently been used to produce spatially explicit predictions of reef fish community variables, our study illustrates the possibility of doing so with satellite data. The ability to use satellite data may bring the cost of creating such maps within the reach of both spatial ecology researchers and the wide range of organizations involved in marine spatial planning.     

Monitoring by Remote Sensing of Habitat Quality Changes in Xianyang City

With the accelerating process of society，economy and urbanization，land use and landscape changes have gradually become important to make effects on regional habitat quality.It is necessary to further investigate those two effects，the result of which can provide a scientific basis for regional habitat conservation and reasonable utilization of land，then will be of great importance in habitat protection and development of the region.In this paper，the study area was located in Xianyang city，Shaanxi Province，which had the frequent human activities and obvious land use changes.Based on the classification of land use data which interpreted by remote sensing，supported by ArcGIS software，the land use transfer matrix of Xianyang from 2000 to 2010 was analyzed.Landscape metrics were calculated by the Fragstats software，which represented for the landscape pattern changes and spatial characteristics.The InVEST model was selected to evaluate habitat quality in study area.The habitat quality changes was monitored.The results indicate that the integrated land use dynamics of Xianyang city is 2.34%，and the changes of land use rate is slow.The main transition from cultivated land and grassland to forest and construction land，which cause the area of first two land use types reduced，and the latter two types increased.The degree of fragmentation and the complexity of structure in landscape are higher than before.Habitat quality improved slightly，and its overall spatial pattern is that central and north areas are relatively higher than the south.Area percentage of excellent，good and poor grades increase，while the habitat quality of medium grade significantly decline.Among 14 districts and counties of habitat quality in Xianyang city，Xunyi County is the best，and the improved magnitude of Liquan county is the most significant.The main driving force of habitat quality change is the transform of land use pattern.Therefore，the relevant departments of Xianyang city should continue to implement the ecological protection measures，and increase the intensity about the protection and management of environment.Thereby to promote the coordinated sustainable development of land use and habitats.     

Linking spatial patterns of bird and butterfly species richness with Landsat TM derived NDVI

The ability to predict spatial patterns of species richness using a few easily measured environmental variables would facilitate timely evaluation of potential impacts of anthropogenic and natural disturbances on biodiversity and ecosystem functions. Two common hypotheses maintain that faunal species richness can be explained in part by either local vegetation heterogeneity or primary productivity. Although remote sensing has long been identified as a potentially powerful source of information on the latter, its principal application to biodiversity studies has been to develop classified vegetation maps at relatively coarse resolution, which then have been used to estimate animal diversity. Although classification schemes can be delineated on the basis of species composition of plants, these schemes generally do not provide information on primary productivity. Furthermore, the classification procedure is a time- and labour-intensive process, yielding results with limited accuracy. To meet decision-making needs and to develop land management strategies, more efficient methods of generating information on the spatial distribution of faunal diversity are needed. This article reports on the potential of predicting species richness using single-date Normalized Difference Vegetation Index (NDVI) derived from Landsat Thematic Mapper (TM). We use NDVI as an indicator of vegetation productivity, and examine the relationship of three measures of NDVI—mean, maximum, and standard deviation—with patterns of bird and butterfly species richness at various spatial scales. Results indicate a positive correlation, but with no definitive functional form, between species richness and productivity. The strongest relationships between species richness of birds and NDVI were observed at larger sampling grains and extent, where each of the three NDVI measures explained more than 50% of the variation in species richness. The relationship between species richness of butterflies and NDVI was strongest over smaller grains. Results suggest that measures of NDVI are an alternative approach for explaining the spatial variability of species richness of birds and butterflies.     

Measurement of habitat predictor variables for organism-habitat models using remote sensing and image segmentation

Robust predictive models of the effects of habitat change on species abundance over large geographical areas are a fundamental gap in our understanding of population distributions, yet are urgently required by conservation practitioners. Predictive models based on underpinning relationships between environmental predictors and the individual organism are likely to require measurement of spatially fine-grained predictor variables. Further, models must show spatial generality if they are to be used to predict the consequences of habitat change over large geographical areas. Remote sensing techniques using airborne scanning laser altimetry (LiDAR) and high resolution multi-spectral imagery allow spatially fine-grained predictor variables to be measured over large geographical areas and thus facilitate testing of the spatial generality of organism-habitat models. These techniques are considered using the skylark as an example species. A range image segmentation system for LiDAR data is described which allows measurement of skylark habitat predictor variables such as within-field vegetation height, boundary height and shape for individual fields within the LiDAR image. Additional variables such as field vegetation type and fractional vegetation ground cover may be obtained from co-registered multi-spectral data. These techniques could have wide application in testing the generality of relationships between populations and habitats, and in ecological monitoring of change in habitat structures and the associated effects on wildlife, over large geographical areas.     

Composition versus physiognomy of vegetation as predictors of bird assemblages: The role of lidar

Whether diversity and composition of avian communities is determined primarily by responses of species to the floristic composition or to the structural characteristics of habitats has been an ongoing debate, at least since the publication of MacArthur and MacArthur (1961). This debate, however, has been hampered by two problems: 1) it is notoriously time consuming to measure the physiognomy of habitat, particularly in forests, and 2) rigorous statistical methods to predict the composition of bird assemblages from assemblages of plants have not been available. Here we use airborne laser scanning (lidar) to measure the habitat (vegetation) structure of a montane forest across large spatial extents with a very fine grain. Furthermore, we use predictive co-correspondence and canonical correspondence analyses to predict the composition of bird communities from the composition and structure of another community (i.e. plants). By using these new techniques, we show that the physiognomy of the vegetation is a significantly more powerful predictor of the composition of bird assemblages than plant species composition in the field and as well in the shrub/tree layer, both on a level of p < 0.001. Our results demonstrate that ecologists should consider remote sensing as a tool to improve the understanding of the variation of bird assemblages in space and time. Particularly in complex habitats, such as forests, lidar is a valuable and comparatively inexpensive tool to characterize the structure of the canopy even across large and rough terrain.     

Assessment of land surface temperature in relation to landscape metrics and fractional vegetation cover in an urban/peri-urban region using Landsat data

Land surface temperature (LST) is essentially considered to be one of the most important indicators used for assessment of the urban thermal environment. It is quite evident that land-use/land-cover (LULC) and landscape patterns have ecological implications at varying spatial scales, which in turn influence the distribution of habitat and material/energy fluxes in the landscape. This article attempts to quantitatively analyse the complex interrelationships between urban LST and LULC landscape patterns with the purpose of elucidating their relation to landscape processes. The study employed an integrated approach involving remote-sensing, geographic information system (GIS), and landscape ecology techniques on bi-temporal Landsat Thematic Mapper images of Southwestern Sydney metropolitan region and the surrounding fringe, taken at approximately the same time of the year in July 1993 and July 2006. First, the LULC categories and LST were extracted from the bi-temporal images. The LST distribution and changes and LST of the LULC categories were then quantitatively analysed using landscape metrics and LST zones. The results show that large differences in temperature existed in even a single LULC category, except for variations between different LULC categories. In each LST zone, the regressive function of LST with fractional vegetation cover (FVC) indicated a significant relationship between LST and FVC. Landscape metrics of LULC categories in each zone in relation to the other zones showed changing patterns between 1993 and 2006. This study also illustrates that a method integrating retrieval of LST and FVC from remote-sensing images combined with landscape metrics provides a novel and feasible way to describe the spatial distribution and temporal variation in urban thermal patterns and associated LULC conditions in a quantitative manner.     

Predicting plant diversity based on remote sensing products in the semi-arid region of Inner Mongolia

Changes in species composition and diversity are the inevitable consequences of climate change, as well as land use and land cover change. Predicting species richness at regional spatial scales using remotely sensed biophysical variables has emerged as a viable mechanism for monitoring species distribution. In this study, we evaluate the utility of MODIS-based productivity (GPP and EVI) and surface water content (NDSVI and LSWI) in predicting species richness in the semi-arid region of Inner Mongolia, China. We found that these metrics correlated well with plant species richness and could be used in biome- and life form-specific models. The relationships were evaluated on the basis of county-level data recorded from the Flora of Inner Mongolia, stratified by administrative (i.e., counties), biome boundaries (desert, grassland, and forest), and grouped by life forms (trees, grasses, bulbs, annuals and shrubs). The predictor variables included: the annual, mean, maximum, seasonal midpoint (EVI_mid), standard deviation of MODIS-derived GPP, EVI, LSWI and NDSVI. The regional pattern of species richness correlated with GPP_SD (R² = 0.27), which was also the best predictor for bulbs, perennial herbs and shrubs (R² = 0.36, 0.29 and 0.40, respectively). The predictive power of models improved when counties with > 50% of cropland were excluded from the analysis, where the seasonal dynamics of productivity and species richness deviate patterns in natural systems. When stratified by biome, GPP_SD remained the best predictor of species richness in grasslands (R² = 0.30), whereas the most variability was explained by NDSVI_max in forests (R² = 0.26), and LSWI_avg in deserts (R² = 0.61). The results demonstrated that biophysical estimates of productivity and water content can be used to predict plant species richness at the regional and biome levels.     

An evolutionary patch pattern approach for texture discrimination

A new evolutionary approach is presented, based on implicit pattern–process relationships. For implementing this approach, any gray level texture image is decomposed into a progressive sequence of binary patch patterns that describe a process of change from background to foreground domination. Each of the binary patterns throughout these sequences is parameterized, using several metrics that describe, for example, its fragmentation level, both for the background (e.g., white) and foreground (e.g., black) patch patterns. Any texture type is then assumed to have a unique evolutionary path represented by a distinctive region in the feature space of metrics characterizing these patterns and their change. Application of hierarchical clustering based on a few (3 or 4) metrics representing characteristic stages in the patterns’ change process allowed us to accurately discriminate between 50 samples of 10 Brodatz texture types.