ebook img

Hurricane Climatology: A Modern Statistical Guide Using R PDF

542 Pages·2013·13.45 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Hurricane Climatology: A Modern Statistical Guide Using R

H U R R I C A N E C L I M AT O L O G Y A Modern Statistical Guide Using R JAMESB.ELSNER&THOMASH.JAGGER TheFloridaStateUniversity Climatek NewYork Oxford OXFORDUNIVERSITYPRESS 2012 1 Preface “Thegoalistoprovideanalyticaltoolsthatwilllaststudents alifetime.” —EdwardTufte Ahurricaneisnature’smostdestructivestorm. Violentwind,flood- ingrain,andpowerfulsurgeposehazardstolifeandproperty. Changes inhurricaneactivitycouldhavesignificantsocietalconsequences.Hurri- canesarealreadytheleadingsourceofinsuredlossesfromnaturalcatas- trophesworldwide. Most of what we know about hurricanes comes from past storms. Hurricane climatology is the study of hurricanes as a collection of past events. It answers questions about when, where, and how often. This book is an argument that there is much more to learn. The goal is to show you how to analyze, model, and predict hurricane climate using data. Itshowsyouhowtocreatestatisticalmodelsfromhurricanedata thatareaccessibleandexplanatory. The book is didactic. It teaches you how to learn about hurricanes from data. It uses statistics. Statistics is the science of organizing and interpretingdata. StatisticshelpsyouunderstandandpredictandtheR programminglanguagehelpsyoudostatistics.Thetextiswrittenaround codethatwhencopiedtoanRsessionwillreproducethegraphs,tables, and maps. The approach is different from other books that use R. It i focusesonasingletopicandshowsyouhowtomakeuseof Rtobetter understandthetopic. The first five chapters provide background material on using R and doingstatistics.Thismaterialisappropriateforanundergraduatecourse onstatisticalmethodsintheenvironmentalsciences.Chapter6presents details on the data sets that are used in the later chapters. Chapters 7, 8,and 9 layout the buildingblocksof models forhurricaneclimate re- search.Thismaterialisappropriateforgraduatelevelcoursesinclimatol- ogy. Chapters10–13giveexamplesfromourmorerecentresearchthat couldbeusedinaseminaronmethodsandmodelsforhurricaneclimate analysisandprediction. The book benefited from research conducted with our students in- cluding Robert Hodges, Jill Trepanier and Kelsey Scheitlin. Editorial assistance came from Laura Michaels with additional help from Sarah Strazzo. Our sincere thanks go to the R Core Development Team for maintainingthissoftwareandtothemanypackageauthorswhoenrich thesoftwareenvironment.WethankRichardMurnaneandTonyKnap of the Risk Prediction Initiative for their unwavering support over the years. GratitudegoestoSvetlaforherwryhumorthroughout. JamesB.Elsner&ThomasH.Jagger Tallahassee,Florida March2012 ii Contents Preface i Contents iii ListofFigures xii ListofTables xviii I Software,Statistics,andData 1 1 Hurricanes,Climate,andStatistics 1 1.1 Hurricanes . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3 Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.4 R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.5 Organization . . . . . . . . . . . . . . . . . . . . . . . . 9 2 RTutorial 12 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.1 WhatisR? . . . . . . . . . . . . . . . . . . . . . 13 2.1.2 GetR . . . . . . . . . . . . . . . . . . . . . . . . 14 2.1.3 Packages . . . . . . . . . . . . . . . . . . . . . . 15 iii 2.1.4 Calculator . . . . . . . . . . . . . . . . . . . . . 16 2.1.5 Functions . . . . . . . . . . . . . . . . . . . . . 17 2.1.6 Warningsanderrors . . . . . . . . . . . . . . . 18 2.1.7 Assignments . . . . . . . . . . . . . . . . . . . . 18 2.1.8 Help . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2 DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.2.1 SmallAmounts . . . . . . . . . . . . . . . . . . 20 2.2.2 Functions . . . . . . . . . . . . . . . . . . . . . 22 2.2.3 Vectors . . . . . . . . . . . . . . . . . . . . . . 23 2.2.4 StructuredData . . . . . . . . . . . . . . . . . . 27 2.2.5 Logic . . . . . . . . . . . . . . . . . . . . . . . . 28 2.2.6 Imports. . . . . . . . . . . . . . . . . . . . . . . 30 2.3 TABLESANDPLOTS . . . . . . . . . . . . . . . . . 34 2.3.1 TablesandSummaries . . . . . . . . . . . . . . 34 2.3.2 Quantiles. . . . . . . . . . . . . . . . . . . . . . 36 2.3.3 Plots . . . . . . . . . . . . . . . . . . . . . . . . 37 BarPlots . . . . . . . . . . . . . . . . . . . . . . 38 ScatterPlots . . . . . . . . . . . . . . . . . . . . 39 2.4 Rfunctionsusedinthischapter . . . . . . . . . . . . . 42 3 ClassicalStatistics 44 3.1 DescriptiveStatistics . . . . . . . . . . . . . . . . . . . 44 3.1.1 Mean,median,andmaximum . . . . . . . . . . 45 3.1.2 Quantiles. . . . . . . . . . . . . . . . . . . . . . 48 3.1.3 Missingvalues . . . . . . . . . . . . . . . . . . . 49 3.2 ProbabilityandDistributions . . . . . . . . . . . . . . . 50 3.2.1 Randomsamples . . . . . . . . . . . . . . . . . 50 3.2.2 Combinatorics . . . . . . . . . . . . . . . . . . . 52 3.2.3 Discretedistributions . . . . . . . . . . . . . . . 53 3.2.4 Continuousdistributions . . . . . . . . . . . . . 55 3.2.5 Distributions . . . . . . . . . . . . . . . . . . . 57 3.2.6 Densities . . . . . . . . . . . . . . . . . . . . . . 57 3.2.7 Cumulativedistributionfunctions . . . . . . . . 59 3.2.8 Quantilefunctions . . . . . . . . . . . . . . . . 61 3.2.9 Randomnumbers . . . . . . . . . . . . . . . . . 62 iv 3.3 One-SampleTests . . . . . . . . . . . . . . . . . . . . . 64 3.4 WilcoxonSigned-RankTest . . . . . . . . . . . . . . . 71 3.5 Two-SampleTests . . . . . . . . . . . . . . . . . . . . . 73 3.6 StatisticalFormula . . . . . . . . . . . . . . . . . . . . . 76 3.7 CompareVariances . . . . . . . . . . . . . . . . . . . . 78 3.8 Two-SampleWilcoxonTest . . . . . . . . . . . . . . . . 79 3.9 Correlation . . . . . . . . . . . . . . . . . . . . . . . . . 80 3.9.1 Pearson’sproduct-momentcorrelation. . . . . . 81 3.9.2 Spearman’srankandKendall’scorrelation. . . . 84 3.9.3 Bootstrapconfidenceintervals . . . . . . . . . . 86 3.9.4 Causation . . . . . . . . . . . . . . . . . . . . . 87 3.10 LinearRegression . . . . . . . . . . . . . . . . . . . . . 87 3.11 MultipleLinearRegression . . . . . . . . . . . . . . . . 96 3.11.1 Predictorchoice . . . . . . . . . . . . . . . . . . 101 3.11.2 Crossvalidation . . . . . . . . . . . . . . . . . . 103 4 BayesianStatistics 104 4.1 LearningAbouttheProportionofLandfalls. . . . . . . 104 4.2 Inference . . . . . . . . . . . . . . . . . . . . . . . . . . 110 4.3 CredibleInterval . . . . . . . . . . . . . . . . . . . . . . 111 4.4 PredictiveDensity . . . . . . . . . . . . . . . . . . . . . 113 4.5 IsBayesRuleNeeded? . . . . . . . . . . . . . . . . . . . 116 4.6 BayesianComputation . . . . . . . . . . . . . . . . . . 117 4.6.1 Time-to-Acceptance . . . . . . . . . . . . . . . 118 4.6.2 MarkovchainMonteCarloapproach . . . . . . 124 4.6.3 JAGS . . . . . . . . . . . . . . . . . . . . . . . . 125 4.6.4 WinBUGS . . . . . . . . . . . . . . . . . . . . . 130 5 GraphsandMaps 136 5.1 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 5.1.1 Boxplot . . . . . . . . . . . . . . . . . . . . . . 137 5.1.2 Histogram . . . . . . . . . . . . . . . . . . . . . 140 5.1.3 Densityplot . . . . . . . . . . . . . . . . . . . . 142 5.1.4 Q-Qplot . . . . . . . . . . . . . . . . . . . . . . 145 5.1.5 Scatterplot . . . . . . . . . . . . . . . . . . . . 147 v 5.1.6 Conditionalscatterplot . . . . . . . . . . . . . . 150 5.2 Timeseries . . . . . . . . . . . . . . . . . . . . . . . . . 152 5.2.1 Time-seriesgraph . . . . . . . . . . . . . . . . . 153 5.2.2 Autocorrelation . . . . . . . . . . . . . . . . . . 153 5.2.3 Datesandtimes . . . . . . . . . . . . . . . . . . 156 5.3 Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 5.3.1 Boundaries . . . . . . . . . . . . . . . . . . . . . 159 5.3.2 Datatypes . . . . . . . . . . . . . . . . . . . . . 162 Pointdata . . . . . . . . . . . . . . . . . . . . . 162 Arealdata . . . . . . . . . . . . . . . . . . . . . 168 Fielddata . . . . . . . . . . . . . . . . . . . . . 171 5.4 CoordinateReferenceSystems . . . . . . . . . . . . . . 174 5.5 Export . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 5.6 OtherGraphicPackages . . . . . . . . . . . . . . . . . . 180 5.6.1 lattice . . . . . . . . . . . . . . . . . . . . . . . . 180 5.6.2 ggplot2 . . . . . . . . . . . . . . . . . . . . . . . 181 5.6.3 ggmap . . . . . . . . . . . . . . . . . . . . . . . 185 6 DataSets 187 6.1 Best-Tracks . . . . . . . . . . . . . . . . . . . . . . . . . 187 6.1.1 Description . . . . . . . . . . . . . . . . . . . . 188 6.1.2 Import . . . . . . . . . . . . . . . . . . . . . . . 190 6.1.3 Intensification . . . . . . . . . . . . . . . . . . . 193 6.1.4 Interpolation . . . . . . . . . . . . . . . . . . . . 194 6.1.5 Regionalactivity . . . . . . . . . . . . . . . . . . 197 6.1.6 Lifetimemaximumintensity . . . . . . . . . . . 198 6.1.7 Regionalmaximumintensity . . . . . . . . . . . 199 6.1.8 Tracksbylocation . . . . . . . . . . . . . . . . . 201 6.1.9 Attributesbylocation . . . . . . . . . . . . . . . 204 6.2 AnnualAggregation . . . . . . . . . . . . . . . . . . . . 206 6.2.1 Annualcyclonecounts . . . . . . . . . . . . . . 206 6.2.2 Environmentalvariables . . . . . . . . . . . . . 208 6.3 CoastalCountyWinds . . . . . . . . . . . . . . . . . . 215 6.3.1 Description . . . . . . . . . . . . . . . . . . . . 215 6.3.2 Countsandmagnitudes . . . . . . . . . . . . . . 217 vi 6.4 NetCDFFiles . . . . . . . . . . . . . . . . . . . . . . . 219 II ModelsandMethods 223 7 FrequencyModels 224 7.1 Counts . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 7.1.1 Poissonprocess . . . . . . . . . . . . . . . . . . 227 7.1.2 InhomogeneousPoissonprocess . . . . . . . . . 228 7.2 EnvironmentalVariables . . . . . . . . . . . . . . . . . 231 7.3 BivariateRelationships . . . . . . . . . . . . . . . . . . 232 7.4 PoissonRegression . . . . . . . . . . . . . . . . . . . . 233 7.4.1 Limitationoflinearregression . . . . . . . . . . 234 7.4.2 Poissonregressionequation . . . . . . . . . . . 235 7.4.3 Methodofmaximumlikelihood . . . . . . . . . 236 7.4.4 Modelfit . . . . . . . . . . . . . . . . . . . . . . 237 7.4.5 Interpretation . . . . . . . . . . . . . . . . . . . 238 7.5 ModelPredictions . . . . . . . . . . . . . . . . . . . . . 240 7.6 ForecastSkill . . . . . . . . . . . . . . . . . . . . . . . . 242 7.6.1 Metrics . . . . . . . . . . . . . . . . . . . . . . . 242 7.6.2 Crossvalidation . . . . . . . . . . . . . . . . . . 243 7.7 NonlinearRegressionStructure . . . . . . . . . . . . . 245 7.8 Zero-InflatedCountModel . . . . . . . . . . . . . . . . 248 7.9 MachineLearning . . . . . . . . . . . . . . . . . . . . . 252 7.10 LogisticRegression . . . . . . . . . . . . . . . . . . . . 255 7.10.1 Exploratoryanalysis . . . . . . . . . . . . . . . . 257 7.10.2 Logitandlogisticfunctions . . . . . . . . . . . . 259 7.10.3 Fitandinterpretation . . . . . . . . . . . . . . . 260 7.10.4 Prediction . . . . . . . . . . . . . . . . . . . . . 262 7.10.5 Fitandadequacy . . . . . . . . . . . . . . . . . 264 7.10.6 Receiveroperatingcharacteristics . . . . . . . . 266 8 IntensityModels 270 8.1 LifetimeHighestIntensity . . . . . . . . . . . . . . . . 270 8.1.1 Exploratoryanalysis . . . . . . . . . . . . . . . . 271 vii

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.