https://orcid.org/0000-0003-3901-4453
Figures
Abstract
The reciprocal relationships between ecological process and landscape pattern are fundamental to landscape ecology. Landscape ecologists traditionally use raster maps portraying classified features such as land use or land cover categories, and metrics suggested by the patch-corridor-matrix conceptual model of pattern. Less attention has been given to the landscape gradient conceptual model and raster maps portraying numeric features such as greenness or percent vegetation cover. We introduce the open-source tool GraySpatCon to calculate and map a variety of landscape pattern metrics from both conceptual models using either categorical or numeric maps. The 51 metrics, drawn mostly from the landscape ecology and image processing literatures, are calculated from the frequencies of input pixel values and/or the pixel value adjacencies in an analysis region. GraySpatCon conducts either a moving window analysis which produces a continuous map of a pattern metric, or a global analysis which produces a single metric value. We describe an implementation in the GuidosToolbox desktop application which allows novice users to interactively explore GraySpatCon functionality. In the R desktop environment, we demonstrate several metrics using an example map of percent tree cover and illustrate a multi-scale moving window analysis to identify scale domains. Comparisons of computational efficiency indicate a substantial GraySpatCon advantage over related software in the R environment.
Citation: Riitters K, Vogt P (2023) Mapping landscape ecological patterns using numeric and categorical maps. PLoS ONE 18(11): e0291697. https://doi.org/10.1371/journal.pone.0291697
Editor: Lalit Kumar Sharma, Zoological Survey of India, INDIA
Received: May 2, 2023; Accepted: August 25, 2023; Published: November 15, 2023
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The reciprocal relationships between ecological process and landscape pattern are fundamental to landscape ecology [1–3]. Landscape pattern metrics quantify the composition and configuration of the elements comprising a landscape [4, 5]. There is a long history of using quantitative and qualitative metrics to examine pattern-process relationships from various perspectives [6–8]. While process-specific ecological interpretation of individual metrics often depends upon local circumstances [c.f. 9, 10], some less-specific yet integrative metrics are commonly applied in the context of biodiversity conservation at international scale [11–13]. Recent reviews demonstrate that research is needed to improve the science and practice of landscape pattern measurement [14–16].
The quantitative analysis of landscape patterns has co-evolved with perspectives about which patterns are important to know about, in turn driven partly by the data and software available to operationalize those perspectives [17]. Early perspectives were dominated by the patch-corridor-matrix conceptual model [18] and spawned a generation of software for analysis of categorical maps [4] as exemplified by FRAGSTATS [19] and landscapemetrics [20]. The advent of maps portraying numeric data (e.g., surface maps of greenness or percent vegetation cover) has contributed to the growing popularity of an alternate conceptual model known as the landscape gradient [21] and recognition that a new generation of pattern analysis is required [22].
While some early landscape ecology software provided limited analysis of numeric data [23], recent ecological research with numeric data has typically used proprietary software from the fields of surface metrology [22, 24–26] and image processing [27–32]. Among the open-source packages available for the R desktop environment, patch-corridor-matrix metrics are available in landscapemetrics [20], surface metrology metrics are available in GEODIV [33], and image texture metrics are available in glcm [34] and fastGLCM [35]. However, none of those tools were developed for both categorical and numeric input data. Furthermore, integrated software systems designed for remote sensing (e.g., Earth Engine [36]), medical imaging (e.g., 3D Slicer [37]), or geographic information systems (e.g., GRASS [38] and QGIS [39]) may provide some relevant procedures but may also be less accessible to non-specialists in those fields.
We developed the open-source C program GraySpatCon (Gray-scale Spatial Convolution, version 1.1.1) to support analysis of a wide range of landscape pattern metrics used by landscape ecologists, including metrics applicable to categorical or numeric raster input data. Despite the popularity of the patch-corridor-matrix conceptual model, its focus on metrics which describe discrete patches in fixed-area landscapes is inconsistent with the landscape gradient conceptual model, in which landscape pattern is a spatially continuous property of a landscape. Furthermore, in contrast to patch-level metrics, pixel-level measures are arguably the fundamental metrics of landscape pattern because patch-level metrics can often be estimated from pixel-level measures [40]. The overall objective of our project is to stimulate and facilitate the application of landscape gradient data in ecological research. The objectives of this paper are to describe the functionality of GraySpatCon and its implementation within several popular computing environments, to illustrate a multi-scale analysis using a selected landscape gradient metric, and to compare computational efficiency with several software alternatives.
Methods
The 51 GraySpatCon metrics are calculated from frequencies of input pixel values and/or pixel value adjacencies, and GraySpatCon is designed primarily for moving window analyses which produce continuous maps of pattern metrics. The moving window algorithm moves an analysis window across the input map, one pixel at a time, accumulating and discarding information along the way. Based on the pixel values and/or adjacencies in the window at a given pixel location, a metric is calculated and assigned to that location on the output map. Thus, the output pixel value codes the landscape pattern context of that pixel location, and the spatial resolution of the input map is preserved. The spatial scale of a moving window analysis is defined by the size of the window; GraySpatCon optionally calculates metrics for a window defined as the entire map extent.
To illustrate basic concepts, a 4 x 4 map contains 16 pixels and 24 adjacencies (Fig 1). Depending on the metric, the moving window algorithm tracks the changing pixel values in a window (summarized in a frequency distribution) and/or the changing pixel value adjacencies in a window (summarized in an adjacency matrix, also known as a co-occurrence matrix). Within an adjacency matrix, the rows and columns indicate the pixel values that are adjacent, and the elements of the matrix indicate the frequencies of each type of adjacency (Fig 2). In GraySpatCon, adjacency matrices are constructed by counting each adjacency in a window once to construct an ordered adjacency matrix; an unordered adjacency matrix is constructed by collapsing the ordered matrix across the main diagonal (Fig 2). Where appropriate, GraySpatCon provides two versions of metrics reflecting the difference between ordered and unordered adjacencies.
Example of a 4 pixel x 4 pixel map containing 16 pixels (left) and 24 adjacencies (right).
Illustration of ordered and unordered adjacency matrices for a 4 pixel x 4 pixel map containing nominal or ordinal data (top). Using a 2-neighbor rule defined as “one pixel below or one pixel to the right,” adjacencies are tabulated to form an ordered adjacency matrix (middle); an unordered adjacency matrix is formed by collapsing the ordered matrix across the main diagonal (bottom). The notation follows that in Table 1. Proportions are rounded to two decimal places.
Using the notation in Table 1, the GraySpatCon metrics are defined in Table 2 where the metric numbers and names follow the nomenclature of the GuidosToolbox (GTB) and GuidosToolbox Workbench (GWB) applications (see below). For consistency with common practice in the image processing literature, the metric definitions assume that numeric data are quantized (“binned”) as integer values in the range [0, 100]; nominal input data can take on any values in that range. Because quantized numeric data are ordinal and therefore categorical, all the metrics can be calculated for a given input map. However, not all metrics are meaningful for all types of input data (Table 2); for example, the correlation metric is meaningless when using nominal data, and the landscape mosaic metric is meaningless when using numeric data. The development philosophy is to provide a generic tool that is potentially applicable to a wide range of investigations. We assume knowledgeable users will select appropriate metrics for a specific circumstance and interpret the results in that context. We recognize the algebraic and geometric correlations among many of the metrics, and we do not advocate mindless calculation or application of them.
Some of the metric definitions in Table 2 differ slightly from other published definitions. This is due partly to differences in the handling of input pixels which are coded as zero (GraySpatCon optionally omits those pixels from calculations), and in the method used to construct adjacency matrices (e.g., ordered, unordered, or symmetric). In addition, certain metrics have the same name in the landscape ecology and image processing literature when different formulas are used, or different names when the formulas are the same.
The default output precision in GraySpatCon is 32-bit floating point, with optional conversion to 8-bit integer precision. GraySpatCon optionally performs a global analysis using the entire extent of the input data area, in which case the output is a text file instead of a map. The moving window algorithm in GraySpatCon is parallelized using OpenMP and by default, will use all cores available in the operating system. The GraySpatCon memory requirement is approximately five times the number of pixels (bytes) in the input map (e.g., ~20 MB RAM for a map of size 2000 x 2000 pixels).
While it is possible for a user to compile and execute a stand-alone application (S1 File), we strongly suggest to first use GraySpatCon via its implementation in the free and open-source image analysis software application GTB (GuidosToolbox) [54]. GTB contains a wide variety of generic raster image processing routines, which are packaged into an interactive desktop application for either Linux, macOS, or MS-Windows. In GTB, the authors have implemented GraySpatCon with a dedicated popup window (Fig 3). This GUI-interface is designed to provide the most intuitive and user-friendly link to GraySpatCon, facilitating the correct interaction and selection of the GraySpatCon parameter settings. Batch-mode, or automatic processing of a series of images, is also available in GTB. The implementation of GraySpatCon in GTB is described further in the GraySpatCon Guide (S1 File), which also contains important usage information and general instructions for a GraySpatCon stand-alone application. For workflow applications in a Linux desktop or server environment, GraySpatCon is also implemented as the GWB_GSC module in the free and open-source software application GWB (GuidosToolbox Workbench) [55]. If computer memory is limited, large images can be processed efficiently by using the GWB_SPLITLUMP module.
Results and discussion
Examples
We illustrate the functionality of GraySpatCon with five landscape pattern metrics applied to a map of percent tree cover in the conterminous United States [56] resampled to a resolution of 2430m (Fig 4 and S2 File). The five metrics include first-order (mean, range) and second-order (correlation, k-contagion) metrics, and a second-order metric targeted at adjacencies involving a specific input pixel value (focal adjacency_1). While a visual comparison of the output maps may suggest the five metrics capture different aspects of the pattern of percent tree cover, the choice of metrics in a specific study naturally depends on the circumstances and the results and interpretations will vary accordingly. Ultimately the perception of pattern depends on the observer. For example, investigations of habitat patterns for different species should use appropriate habitat data and metrics which describe the aspects of pattern that are believed to be important from the perspective of each species [4]. Since all potential observers cannot be known in advance, GraySpatCon is designed to quantify and map many pattern metrics in a consistent way such that comparisons among different studies may be facilitated.
Five landscape pattern metrics were mapped using a window size of 31x31 pixels (5675 km2). The panels show the input data (a) [56] and output data (b–f) for five metrics (respectively, metric numbers 1, 44, 27, 22, 49 in Table 2). Darker colors indicate larger values, ecoregion boundaries [57] are shown for comparisons, and the data range in each panel is indicated.
It is also well known that the perceived pattern depends on the observation scale. In ecology this is often translated to a requirement of selecting an observation scale which is appropriate for a specific ecological process or species [4]. In a broader view, ecologists are also interested in how pattern per se changes with observation scale in different landscapes; for example, “scalograms” [58] may suggest “scale domains” [59] over which pattern-process relationships may be stable. The “scale parameter” In GraySpatCon is only one aspect of observation scale–the window size which defines the spatial extent of the analysis. Fig 5 illustrates the effect of changing the window size on the Gini-Simpson diversity metric as applied to the map of percent tree cover. An example of a scalogram was constructed using 21 window sizes from 5x5 pixels to 45x45 pixels and plotting the maximum metric value (over all windows) versus window size (Fig 6). The stability of the maximum metric value in window sizes from 11x11 to 25x25 suggests a plausible scale domain, which could be tested for significance and subsequently compared to an ecological process which was measured at the same 21 observation scales.
Illustration of changing the observation scale (window size) for the Gini-Simpson diversity metric (Table 2; metric 10) as applied to a map of percent tree cover (a) [56]. The diversity of percent tree cover was mapped for the three indicated window sizes (b, c, d) corresponding to window areas of 53 km2, 714 km2, and 5675 km2). Darker colors indicate larger values, ecoregion boundaries [57] are shown for comparisons, and the data range in each panel is indicated.
The observation scale is indicated by the side length of the moving window, for example an observation scale of 15 corresponds to a window size of 15x15 pixels. The plausible scale domain is a range of observation scales over which the maximum value is relatively stable.
The example in Fig 6 used the maximum metric value over all windows to illustrate a simple scalogram. While beyond the scope of this illustration, the multi-scale results for each location (i.e., each window) could be collectively analyzed to identify and map the locations of prevalent scalograms [60] and identify the scales and locations of transitions from global to local scaling relationships [61]. While the traditional patch-corridor-matrix metrics may be better suited for describing discrete patches, similar multi-scale procedures could be applied to nominal data to describe imputed patches in terms perimeter characteristics, juxtaposition, size, and spatial distribution [62, 63].
Computational efficiency
Computational efficiency is an important aspect of landscape pattern analysis, especially when addressing problems requiring calculations for large or numerous input maps, or for many metrics or window sizes. A full comparison of all GraySpatCon metrics with other software is not possible because no other software implements all the GraySpatCon metrics. We selected a set of comparable metrics for illustration and conducted all trials in the R desktop environment [64] using the map of percent tree cover described above (S2 File). We compared GraySpatCon moving window analyses of four second-order Haralick [43] texture metrics with the R packages glcm [34] and fastGLCM [35] (Table 3). We also compared GraySpatCon global and moving window analyses of one second-order and two first-order metrics with the R package landscapemetrics [20] (Table 4). In the moving window comparisons GraySpatCon was substantially faster than all three alternatives. In the global analysis comparisons, GraySpatCon was substantially faster than landscapemetrics for two of the three metrics tested. While alternate software may be required for metrics not implemented in GraySpatCon, we believe these limited comparisons are a compelling argument to consider using GraySpatCon when there is a choice.
The gray-scale input map (Fig 4A) is 1990 X 1289 pixels, the window size is 31x31 pixels, and the reported time is the mean of three trials.
The gray-scale input image (Fig 4A) is 1990 X 1289 pixels, the window size is 3x3 pixels, and the reported time is the mean of three trials. Multi-core times were not compared because landscapemetrics supports only single-core usage, and for practical reasons the smallest possible window size was used.
Conclusion
Landscape ecologists use a variety of software and metrics to quantify and map landscape patterns using raster data, but there are relatively few open-source software alternatives for conducting many of the analyses commonly used to explore pattern-process relationships, particularly software that supports analysis of numeric input data. We developed GraySpatCon to quantify and map a wide range of landscape pattern metrics using categorical or numeric input data, and our comparisons with alternate software indicated a significant advantage in computational efficiency. We implemented GraySpatCon in the popular GuidosToolbox desktop application so that novice users can interactively explore its capabilities, and in the GuidosToolbox Workbench so that experienced users can easily integrate its capabilities in computationally intensive workflows. Alternatively, binary executable versions of GraySpatCon are available for three popular operating systems for implementation in other desktop applications such as R (S2 File). For developers, the source code for GraySpatCon is distributed on GitHub under CC0 (CC0 1.0 Universal [CC0 1.0] Public Domain Dedication) (S2 File). Through these efforts we hope to stimulate increased attention to application of numeric maps in ecology and related fields, and consistent treatment of nominal and numeric data where appropriate.
Supporting information
S1 File. The Guide describes implementation in the GuidosToolbox application and provides important usage notes.
https://doi.org/10.1371/journal.pone.0291697.s001
(PDF)
S2 File. This archive contains example input maps and R scripts used for examples and comparisons, and instructions to download GraySpatCon source code and binary executable files for 64-bit Linux, macOS, and MS-Windows operating systems.
https://doi.org/10.1371/journal.pone.0291697.s002
(ZIP)
Acknowledgments
Jennifer Costanza and Karen Schleeweis reviewed the R scripts (S2 File). This work was performed in the context of a research agreement between the USDA Forest Service (Memorandum of Understanding #14-MU-11330110-001) and the European Commission Joint Research Centre (Collaborative Research Arrangement #33385).
References
- 1. Troll C. 1971. Landscape ecology (geoecology) and biogeocenology—A terminological study. Geoforum 2:43–46. https://doi.org/10.1016/0016-7185(71)90029-7
- 2.
Forman RTT, Godron M. 1986. Landscape Ecology. Wiley, New York.
- 3. Turner MG. 1989. Landscape ecology: the effect of pattern on process. Annual Review of Ecology and Systematics 20:171–197. https://doi.org/10.1146/annurev.es.20.110189.001131
- 4. Gustafson E. 1998. Quantifying landscape spatial pattern: what is the state of the art? Ecosystems 1:143–156. https://doi.org/10.1007/s100219900011
- 5. Gustafson E. 2019. How has the state-of-the-art for quantification of landscape pattern advanced in the twenty-first century? Landscape Ecology 34:2065–2072.
- 6. Turner MG. 2005. Landscape ecology: what Is the state of the science? Annual Review of Ecology, Evolution, and Systematics 36:319–344. https://doi.org/10.1146/annurev.ecolsys.36.102003.152614
- 7.
Bell S. 2012. Landscape pattern, perception, and process. Routledge.
- 8.
Turner MG, Gardner RH. 2015. Landscape ecology in theory and practice, 2nd edition. Springer, New York. https://doi.org/10.1007/978-1-4939-2794-4
- 9. Fahrig L. 2017. Ecological responses to habitat fragmentation per se. Annual Review of Ecology, Evolution, and Systematics 48:1–23. https://doi.org/10.1146/annurev-ecolsys-110316-022612
- 10. Fletcher RJ, Didham RK, Banks-Leite C, Barlow J, Ewers RM, Rosindell J, et al. 2018. Is habitat fragmentation good for biodiversity? Biological Conservation 226:9–15. https://doi.org/10.1016/j.biocon.2018.07.022
- 11. FAO and UNEP. 2020. The State of the World’s Forests 2020. Forests, biodiversity and people. Rome. https://doi.org/10.4060/ca8642en
- 12.
Forest Europe. 2020. State of Europe’s forests 2020. Ministerial conference on the protection of forests in Europe–FOREST EUROPE. Liaison Unit Bratislava. https://foresteurope.org
- 13.
Maes J, Teller A, Erhard M, Condé S, Vallecillo S, Barredo JI, et al. 2020. Mapping and assessment of ecosystems and their services: an EU ecosystem assessment, Publications Office of the European Union, Luxembourg. https://publications.jrc.ec.europa.eu/repository/handle/JRC120383.
- 14. Kupfer JA. 2012. Landscape ecology and biogeography: rethinking landscape metrics in a post-FRAGSTATS landscape. Progress in Physical Geography 36:400–420. https://doi.org/10.1177/0309133312439594
- 15. Lausch A, Blaschke T, Haase D, Herzog F, Syrbe RU, Tischendorf L, et al. 2015. Understanding and quantifying landscape structure—A review on relevant process characteristics, data models and landscape metrics. Ecological Modelling 295:31–41. https://doi.org/10.1016/j.ecolmodel.2014.08.018
- 16. Frazier AE, Kedron P. 2017. Landscape metrics: past progress and future directions. Current Landscape Ecology Reports 2:63–72. https://doi.org/10.1007/s40823-017-0026-0
- 17. Costanza JK, Riitters K, Vogt P, Wickham J. 2019. Describing and analyzing landscape patterns: where are we now, and where are we going? Landscape Ecology 34:2049–2055. https://doi.org/10.1007/s10980-019-00889-6
- 18. Forman RTT. 1995. Some general principles of landscape and regional ecology. Landscape Ecology 10:133–142. https://doi.org/10.1007/BF00133027
- 19. McGarigal K, Marks BJ. 1995. Spatial pattern analysis program for quantifying landscape structure. Gen. Tech. Rep. PNW-GTR-351. US Department of Agriculture, Forest Service, Pacific Northwest Research Station, 122 pages. https://doi.org/10.2737/PNW-GTR-351
- 20. Hesselbarth MH, Sciaini M, With KA, Wiegand K, Nowosad J. 2019. Landscapemetrics: an open‐source R tool to calculate landscape metrics. Ecography 42:1648–1657.
- 21.
McGarigal K, Cushman SA. 2005. The gradient concept of landscape structure. In: Wiens J, Moss M (eds), Issues and Perspectives in Landscape Ecology. Cambridge University Press, Cambridge, pp 112–119.
- 22. Hoechstetter S, Walz U, Dang LH, Thinh NX. 2008. Effects of topography and surface roughness in analyses of landscape structure–A proposal to modify the existing set of landscape metrics. Landscape Online 3:1–14. https://doi.org/10.3097/LO.200803
- 23. Baker WL, Cai Y. 1992. The r.le programs for multiscale analysis of landscape structure using the GRASS geographical information system. Landscape Ecology 7:291–302. https://doi.org/10.1007/BF00131258
- 24. McGarigal K, Tagil S, Cushman SA. 2009. Surface metrics: an alternative to patch metrics for the quantification of landscape structure. Landscape Ecology 24:433–450. https://doi.org/10.1007/s10980-009-9327-y
- 25. Abdel Moniem HEM, Holland JD. 2013. Habitat connectivity for pollinator beetles using surface metrics. Landscape Ecology 28:1251–1267. https://doi.org/10.1007/s10980-013-9886-9
- 26. Kedron PJ, Frazier AE, Ovando-Montejo GA, Wang J. 2018. Surface metrics for landscape ecology: a comparison of landscape models across ecoregions and scales. Landscape Ecology 33:1489–1504. https://doi.org/10.1007/s10980-018-0685-1
- 27. St-Louis V, Pidgeon AM, Radeloff VC, Clayton MK. 2006. High-resolution image texture as a predictor of bird species richness. Remote Sensing of Environment 105:299–312. https://doi.org/10.1016/j.rse.2006.07.003
- 28. Tuttle EM, Jensen RR, Formica VA, Gonser RA. 2006. Using remote sensing image texture to study habitat use patterns: a case study using the polymorphic white-throated sparrow (Zonotrichia albicollis). Global Ecology and Biogeography 15:349–357. https://doi.org/10.1111/j.1466-822X.2006.00232.x
- 29. St-Louis V, Pidgeon AM, Clayton MK, Locke BA, Bash DW, Radeloff VC. 2009. Satellite image texture and a vegetation index predict avian biodiversity in the Chihuahuan Desert of New Mexico. Ecography 32:468–480. https://doi.org/10.1111/j.1600-0587.2008.05512.x
- 30. Wood EM, Pidgeon AM, Radeloff VC, Keuler NS. 2013. Image texture predicts avian density and species richness. PLoS One, 8(5): e63211. pmid:23675463
- 31. Tuanmu M-N, Jetz W. 2015. Global habitat heterogeneity. Global Ecology and Biogeography 24:1329–1339. https://doi.org/10.1111/geb.12365
- 32. Gallardo‐Cruz JA, Hernández‐Stefanoni JL, Moser D, Martínez‐Yrizar A, Llobet S, Meave JA. 2018. Relating species richness to the structure of continuous landscapes: alternative methodological approaches. Ecosphere 9(5), e02189. https://doi.org/10.1002/ecs2.2189
- 33. Smith AC, Dahlin KM, Record S, Costanza JK, Wilson AM, Zarnetske PL. 2021. The geodiv R package: Tools for calculating gradient surface metrics. Methods in Ecology and Evolution, 12:2094–2100. pmid:35874973
- 34. Zvoleff A. 2020. glcm: Calculate textures from grey-level co-occurrence matrices (GLCMs). R package version 1.6.5. https://CRAN.R-project.org/package=glcm.
- 35. Mouselimis L. 2022. _fastGLCM: Fast gray level co-occurrence matrix computation (GLCM) using R. R package version 1.0.2, <URL: https://CRAN.R-project.org/package=fastGLCM>
- 36. Gorelick N, Hancher M, Dixon M, Ilyushchenko S, Thau D, Moore R. 2017. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment 202:18–27. https://doi.org/10.1016/j.rse.2017.06.031
- 37. Fedorov A, Beichel R, Kalpathy-Cramer J, Finet J, Fillion-Robin , et al. 2012. 3D Slicer as an image computing platform for the quantitative imaging network. Magnetic Resonance Imaging 30:1323–1341. pmid:22770690
- 38.
GRASS Development Team. 2022. Geographic Resources Analysis Support System (GRASS) software, Version 8.2. Open Source Geospatial Foundation. https://grass.osgeo.org.
- 39. QGIS Development Team. 2022. QGIS Geographic Information System. Open Source Geospatial Foundation Project. http://qgis.osgeo.org
- 40. Riitters K. 2019. Pattern metrics for a transdisciplinary landscape ecology. Landscape Ecology 34:2057–2063. https://doi.org/10.1007/s10980-018-0755-4
- 41. Li H, Reynolds JF. 1993. A new contagion index to quantify spatial patterns of landscapes. Landscape Ecology 8:155–162. https://doi.org/10.1007/BF00125347
- 42. Riitters KH O’Neill RV, Wickham JD, Jones KB. 1996. A note on contagion indices for landscape analysis. Landscape Ecology 11:197–202. https://doi.org/10.1007/BF02071810
- 43. Haralick RM, Shanmugam K, Dinstein I. 1973. Textural features for image classification. In: IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-3, no. 6, pp. 610–621. https://doi.org/10.1109/TSMC.1973.4309314
- 44. Riitters KH, O’Neill RV, Hunsaker CT, Wickham JD, Yankee DH, Timmins SP, et al. 1995. A factor analysis of landscape pattern and structure metrics. Landscape Ecology 10:23–39. https://doi.org/10.1007/BF00158551
- 45. Wickham JD, Riitters KH. 1995. Sensitivity of landscape metrics to pixel size. International Journal of Remote Sensing 16:3585–3594. https://doi.org/10.1080/01431169508954647
- 46. Riitters K, Costanza JK, Coulston JW, Vogt P, Schleeweis K. 2023. Interpreting image texture metrics applied to landscape gradient data. Landscape Ecology 38. https://doi.org/10.1007/s10980-023-01705-y
- 47. Riitters KH O’Neill RV, Jones KB. 1997. Assessing habitat suitability at multiple scales: A landscape-level approach. Biological Conservation 81:191–202. https://doi.org/10.1016/S0006-3207(96)00145-0
- 48. Riitters K, Wickham J, O’Neill R, Jones B, Smith E. 2000. Global-scale patterns of forest fragmentation. Conservation Ecology 4(2): 3. http://www.consecol.org/vol4/iss2/art3/
- 49. Wade TG, Riitters KH, Wickham JD, Jones KB. 2003. Distribution and causes of global forest fragmentation. Conservation Ecology 7(2). http://www.consecol.org/vol7/iss2/art7
- 50. Soh L-K, Tsatsoulis C. 1999. Texture analysis of SAR sea ice imagery using gray level co-occurrence matrices. IEEE Transactions on Geoscience and Remote Sensing 37:780–795. https://doi.org/10.1109/36.752194
- 51. Clausi DA. 2002. An analysis of co-occurrence texture statistics as a function of grey level quantization. Canadian Journal of Remote Sensing 28:45–62. https://doi.org/10.5589/m02-004
- 52. Gower JC. 1971. A general coefficient of similarity and some of its properties. Biometrics 27:857–871. https://doi.org/10.2307/2528823
- 53. Conners RW, Trivedi MM, Harlow CA. 1984. Segmentation of a high-resolution urban scene using texture operators. Computer Vision, Graphics, and Image Processing 25:273–310. https://doi.org/10.1016/0734-189X(84)90197-X
- 54. Vogt P, Riitters K. 2017. GuidosToolbox: universal digital image object analysis. European Journal of Remote Sensing 50:352–361. https://doi.org/10.1080/22797254.2017.1330650
- 55. Vogt P, Riitters K, Rambaud P, d’Annunzio R, Lindquist E, Pekkarinen A. 2022. GuidosToolbox Workbench: spatial analysis of raster maps for ecological applications. Ecography, 2022(3), e05864. https://doi.org/10.1111/ecog.05864
- 56.
USGS (United States Geological Survey). 2019. NLCD 2016 tree canopy cover (CONUS). U.S. Geological Survey, Sioux Falls, SD. URL: https://www.mrlc.gov/data/nlcd-2016-usfs-tree-canopy-cover-conus
- 57.
Cleland DT, Freeouf JA, Keys JE, Nowacki G J, Carpenter J, McNab WH. 2007. Ecological subregions: Sections and subsections of the conterminous United States [1:3,500,000] [CD-ROM]. Sloan AM, cartographer. U.S. Department of Agriculture, Forest Service, General Technical Report WO-76, Washington, DC.
- 58. Wu J, Shen W, Sun W, Tueller PT. 2002. Empirical patterns of the effects of changing scale on landscape metrics. Landscape Ecology 17:761–782. https://doi.org/10.1023/A:1022995922992
- 59. Wiens JA. 1989. Spatial scaling in ecology. Functional Ecology 3:385–397.
- 60. Zurlini G, Riitters KH, Zaccarelli N, Petrosillo I. 2007. Patterns of disturbance at multiple scales in real and simulated landscapes. Landscape Ecology 22:705–721. https://doi.org/10.1007/s10980-006-9055-5
- 61. Zaccarelli N, Petrosillo I, Zurlini G, Riitters KH. 2008. Source/sink patterns of disturbance and cross-scale mismatches in a panarchy of social-ecological landscapes. Ecology and Society 13(1). http://www.jstor.org/stable/26267926
- 62. Riitters KH, Costanza JK, Buma B. 2017. Interpreting multiscale domains of tree cover disturbance patterns in North America. Ecological Indicators 80:147–152. https://doi.org/10.1016/j.ecolind.2017.05.022
- 63. Zhang Y, Guindon B. 2017. Detecting and quantifying extended landscape structure with spatial co-occurrence surfaces. Pattern Analysis and Applications 20:519–529. https://doi.org/10.1007/s10044-015-0518-7
- 64.
R Core Team. 2022. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/