{"id":2048,"date":"2016-06-08T21:47:38","date_gmt":"2016-06-09T04:47:38","guid":{"rendered":"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/?p=2048"},"modified":"2016-06-08T23:11:07","modified_gmt":"2016-06-09T06:11:07","slug":"exploratory-analysis-forest-disturbance-clustering-using-kernel-density-morans","status":"publish","type":"post","link":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/2016\/06\/08\/exploratory-analysis-forest-disturbance-clustering-using-kernel-density-morans\/","title":{"rendered":"Exploratory Analysis of Forest Disturbance Distribution through Kernel Density Mapping"},"content":{"rendered":"<p>To begin my analysis of\u00a0forest disturbance patterns, I wanted to get a broad scale understanding of how clear-cuts and partial harvests are distributed throughout\u00a0Willamette National Forest, as well as the land use designations and forest districts they spatially coincide with. Using ArcMap I ran the\u00a0Kernel Density tool\u00a0on cumulative disturbances from 1985 to\u00a02012. This first required converting my disturbance patch data (rasters) into polygons (vectors). Additionally, Kernel Density requires point or polyline inputs, so I also generated disturbance\u00a0centroids from the polygons. Below I outline the function, results and my interpretations:<\/p>\n<p><span style=\"font-weight: 400\"><strong>Kernel Density<\/strong> calculates a magnitude per unit area using a kernel function to fit a smoothly tapered surface to each input feature. Here are the results for two runs of this tool:<\/span><\/p>\n<p><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/disturb_density3.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-2061 alignleft\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/disturb_density3.jpg\" alt=\"disturb_density3\" width=\"637\" height=\"824\" srcset=\"https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density3.jpg 2550w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density3-232x300.jpg 232w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density3-768x994.jpg 768w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density3-791x1024.jpg 791w\" sizes=\"auto, (max-width: 637px) 100vw, 637px\" \/><\/a><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/disturb_density2-1.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-2062\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/disturb_density2-1.jpg\" alt=\"disturb_density2\" width=\"635\" height=\"821\" srcset=\"https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density2-1.jpg 2550w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density2-1-232x300.jpg 232w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density2-1-768x994.jpg 768w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/disturb_density2-1-791x1024.jpg 791w\" sizes=\"auto, (max-width: 635px) 100vw, 635px\" \/><\/a><\/p>\n<p>Below\u00a0are the parameters used for each kernel density output above. In the first run on the left, the output cell size of 250 meters results in a surface that is less smooth than that of the second run on the right, which has a much smaller output cell size. Additionally, in the second run, the optional parameter for\u00a0a search radius was set to\u00a01600\u00a0meters, a value that is roughly double that of the average nearest-neighbor distance between disturbance centroids. This creates a density surface that is more appropriate for the scale of the map (on the right). Both maps give an indication that higher densities of clear-cuts and partial harvests between 1984 and 2012 occurred on Matrix Lands and Non-Forest Service Lands, which are mostly composed of\u00a0private and industrial landowners. These results are not surprising, given that timber harvest is the primary function of Matrix Lands, and that private and industrial landowners are inclined to produce timber.<\/p>\n<p><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/Capture.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-2063\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/Capture.jpg\" alt=\"Capture\" width=\"459\" height=\"265\" \/><\/a><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/Capture3.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-2066\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/Capture3.jpg\" alt=\"Capture3\" width=\"464\" height=\"265\" \/><\/a><\/p>\n<p>*It is important to note that after\u00a0converting the disturbance patches to polygons, I used the Normal QQ Plot and Histogram functions (built in to the Geostatistical\u00a0Analyst extension for ArcMap) to remove outliers using the &#8220;Shape_Area&#8221; attribute field. Polygons of extremely low area (less than eleven 30-meter Landsat pixels) were removed because they likely represent incorrectly classified pixels. Polygons of extremely high area were also removed, because they would not be accurately represented by a single centroid point.<\/p>\n<p>Mapping cumulative clear-cuts and partial harvests over\u00a0the 27 year period between 1985 and 2012 paints an interesting picture, but\u00a0in order to better assess the effects of forest governance on landscape patterns, it is probably more interesting to map disturbances temporally. Using the Definition Query function in ArcMap, I filtered the disturbance data at 5-year intervals, and ran the Kernel Density tool again for each interval:<\/p>\n<p><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/dens_time.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-2155 aligncenter\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/dens_time.jpg\" alt=\"dens_time\" width=\"609\" height=\"733\" srcset=\"https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/dens_time.jpg 2819w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/dens_time-249x300.jpg 249w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/dens_time-768x924.jpg 768w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/dens_time-851x1024.jpg 851w\" sizes=\"auto, (max-width: 609px) 100vw, 609px\" \/><\/a><\/p>\n<p>Bringing in the temporal dimension reveals some interesting changes in the density of forest disturbance that may correspond with significant events in the history of forest governance. For example, 1990 shows very high density of clear-cuts and partial harvests, which indicates a peak in timber harvest activity prior to the May 29, 1991 Dwyer Injunction, which banned timber sales. Following the 1994 Record of Decision for\u00a0the Northwest Forest Plan, the maps for 2000, 2005 and 2010 show the expected\u00a0drop in overall disturbance density, but interesting peaks associated with certain forest districts. For one final run of the Kernel Density tool, I mapped cumulative disturbance during the period of the Dwyer Injunction (1990-1994). The result shows higher disturbance density than expected, but I assume that there is\u00a0lag between the timing of timber sales and when sold forest land is actually harvested. Thus, this map likely shows clear-cuts and harvests on forest lands sold prior to the injunction, but may also be indicative of\u00a0which forest districts are more inclined toward timber production.<\/p>\n<p><a href=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/91_94_dens-1.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-2170 aligncenter\" src=\"http:\/\/blogs.oregonstate.edu\/geo599spatialstatistics\/files\/2016\/06\/91_94_dens-1.jpg\" alt=\"91_94_dens\" width=\"581\" height=\"752\" srcset=\"https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/91_94_dens-1.jpg 2550w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/91_94_dens-1-232x300.jpg 232w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/91_94_dens-1-768x994.jpg 768w, https:\/\/osu-wams-blogs-uploads.s3.amazonaws.com\/blogs.dir\/1572\/files\/2016\/06\/91_94_dens-1-791x1024.jpg 791w\" sizes=\"auto, (max-width: 581px) 100vw, 581px\" \/><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>To begin my analysis of\u00a0forest disturbance patterns, I wanted to get a broad scale understanding of how clear-cuts and partial harvests are distributed throughout\u00a0Willamette National Forest, as well as the land use designations and forest districts they spatially coincide with. Using ArcMap I ran the\u00a0Kernel Density tool\u00a0on cumulative disturbances from 1985 to\u00a02012. This first required&hellip; <a href=\"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/2016\/06\/08\/exploratory-analysis-forest-disturbance-clustering-using-kernel-density-morans\/\">Continue reading <span class=\"meta-nav\">&rarr;<\/span><\/a><\/p>\n","protected":false},"author":7726,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[661527],"tags":[],"class_list":["post-2048","post","type-post","status-publish","format-standard","hentry","category-tutorials-2016"],"_links":{"self":[{"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/posts\/2048","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/users\/7726"}],"replies":[{"embeddable":true,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/comments?post=2048"}],"version-history":[{"count":1,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/posts\/2048\/revisions"}],"predecessor-version":[{"id":2172,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/posts\/2048\/revisions\/2172"}],"wp:attachment":[{"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/media?parent=2048"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/categories?post=2048"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dev.blogs.oregonstate.edu\/geo599spatialstatistics\/wp-json\/wp\/v2\/tags?post=2048"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}