Philippe A relatively simple approach: You could slice the time and prepare kernel density maps for each time period. For each of these time periods, perform K-function at different scales, identify at which scales the clustering is the strongest (using L-function), and use that as an input for kernel density maps. Then you have a sequence of maps you can animate to understand the change over time. You could then estimate whether density of clustering of the animals has a spatial correlation with proximity to buildings. More info here: Delmelle, E. 2009. Point Pattern Analysis. In Kitchin R, Thrift N (eds) International Encyclopedia of Human Geography. 8: 204-211. Oxford: Elsevier. A more complicated approach: You may want to consider using network-based kernels since we know that although animals can move freely, they tend to follow distinct path. Joni Downs and Mark Horner have done some work on NKDE (network kernel density estimation). It is computationally harder (using delaunay triangulation), but the results are more robust. Downs and Horner (2008): Spatially modeling parthways of migratory birds for nature reserve site selection IJGIS Downs and Horner (2008): Effects of point pattern shape on home-range estimate They also have some pdfs you can find on the internet (JA Downs and MW Horner). eric
-- (704) 687 5991 Charlotte, NC 28223 http://www.geoearth.uncc.edu/faculty/edelmel1/ -- ________________________________ From: [email protected] on behalf of Philippe Bouchet Sent: Wed 5/5/2010 01:33 To: [email protected] Subject: AI-GEOSTATS: Point pattern analysis and animal displacement Dear list, I am working on a project which aims to determine whether the construction of industrial facilities near to / in the middle of a migratory path for large marine animals has an effect on the spatial distribution of those migrating animals. My dataset consists of a series of points marking the location (GPS coordinates) of animals sighted during several dedicated aerial surveys over the area (before and after the construction of the industrial platform), and I also know the position of the facilities of course. How can relate the spatial distribution of animals to the presence of the industrial facilities, with the objective of testing whether the animals have been displaced from their normal route ? My initial thoughts on this were to: 1) test for CSR (Complete Spatial Randomness) in the point pattern - if the animals were distributed randomly over the area prior to the implantation of the facilities but now display a clustered or gradient pattern in distribution, this could be indicative of a possible displacement. 2) Construct 2D kernel density estimates, using appropriate functions in R, for each day an aerial survey was carried out - that would enable me to understand how the distribution of animals changes through time.ยจ Is this the right way to go ? Are there other tools / analyses out there that may be more suited to answering my question and that I may not be aware of ? Any ideas or suggestions much welcome and greatly appreciated, Many thanks in advance, Philippe + + To post a message to the list, send it to [email protected] + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu with no subject and "unsubscribe ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
