the r book simulation models

《the r book simulation models》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、26SpatialStatisticsTherearethreekindsofproblemsthatyoumighttacklewithspatialstatistics:
pointprocesses(locationsandspatialpatternsofindividuals);
mapsofacontinuousresponsevariable(kriging);
spatiallyexplicitresponsesaffectedbytheidentity,sizeandproximityofneighbours.26.1PointprocessesTherearet

2、hreebroadclassesofspatialpatternonacontinuumfromcompleteregularity(evenlyspacedhexagonswhereeveryindividualisthesamedistancefromitsnearestneighbour)tocompleteaggregation(alltheindividualsclusteredintoasingleclump):wecalltheseregular,randomandaggregatedpatternsandtheylooklikethis:regularrandoma

3、ggregatedIntheirsimplestform,thedataconsistofsetsofxandycoordinateswithinsomesamplingframesuchasasquareoracircleinwhichtheindividualshavebeenmapped.Thefirstquestionisoftenwhetherthereisanyevidencetoallowrejectionofthenullhypothesisofcompletespatialrandomness(CSR).Inarandompatternthedistribution

4、ofeachindividualiscompletelyindependentofthedistributionofeveryTheRBook,SecondEdition.MichaelJ.Crawley.©2013JohnWiley&Sons,Ltd.Published2013byJohnWiley&Sons,Ltd.826THERBOOKother.Individualsneitherinhibitnorpromoteoneanother.Inaregularpatternindividualsaremorespacedoutthaninarandomone,presumabl

5、ybecauseofsomemechanism(suchascompetition)thateliminatesindividualsthataretooclosetogether.Inanaggregatedpattern,individualaremoreclumpedthaninarandomone,presumablybecauseofsomeprocesssuchasreproductionwithlimiteddispersal,orbecauseofunderlyingspatialheterogeneity(e.g.goodpatchesandbadpatches)

6、.Countsofindividualswithinsampleareas(quadrats)canbeanalysedbycomparingthefrequencydistributionofcountswithaPoissondistributionwiththesamemean.Aggregatedspatialpatterns(inwhichthevarianceisgreaterthanthemean)areoftenwelldescribedbyanegativebinomialdistribu-tionwithaggregationparameterk(seep.31

7、5).Themainproblemwithquadrat-basedcountsisthattheyarehighlyscale-dependent.Thesamespatialpatterncouldappeartoberegularwhenanalysedwithsmallquadrats,aggregatedwhenanalysedwithmedium-sizedquadrats,yetrandomwhenanalysedwithlargequadrats.Di