2012 AGU Chapman Conference on Remote Sensing of the ...

ealistic models of the structure - i.e. to tame the variability -is the existence of some scale by scale regularity, the existenceof regimes which are in some sense scale invariant over wideranges of space-time scales. Over the past thirty years it hasbecome clear that scale invariance is a symmetry principle ofgreat generality; a system can scale invariant even when it ishighly anisotropic so that small and large structure arestatistically related by “zooms” followed by the squashingand rotation of structures. It is also known thatmultiplicative cascades are typically associated withnonlinear scale invariant dynamics. Today, technologicaladvances have permitted observations of precipitation overwide ranges of scales including the extreme small drop scales(using stereophotography) and the extreme large planetaryscales (using satellite borne radar) as well as meteorologicalreanalyses and to continental scale gridded raingageproducts. Below scales of roughly 30 - 50 cm (depending onthe rate) the rain decouples from the turbulence formingstatistically homogeneous “patches”. At larger scales, on thecontrary, the precipitation is strongly coupled with theturbulence so that the liquid water statistics are very nearlythose of passive scalars. Using in situ rain gage networks,ground and satellite radar data and meteorologicalreanalyses, we show that the precipitation has amultiplicative cascade structure up to planetary scales inspace and up to w 5- 10 days in time. At times longer thanw - the lifetime of planetary scale structures – the statisticsof atmospheric fields (including precipitation) becomemuch “calmer”, the spatial degrees of freedom are essentiallyquenched (a “dimensional transition”). We have theemergence of a new “low frequency weather” regime whosestatistics we characterize. But precipitation is stronglynonlinearly coupled to the other atmospheric fields. To becredible, this scale invariant cascade dynamic must apply tothe other fields as well. We outline such a paradigm based ona) advances in the last 25 years in nonlinear dynamics, b) acritical reanalysis of empirical aircraft and vertical sondedata, c) the systematic scale by scale space-time exploitationof high resolution remotely sensed data (TRMM radar andradiances, MTSAT radiances) d) the systematic reanalysis ofthe outputs of numerical models of the atmosphereincluding GFS, GEM models and the ERA40, and the NOAA20th Century reanalyses) and e) a new turbulent model forthe emergence of the climate from “weather” and climatevariability. We discuss applications to the remote sensing ofrain and implications for hydrology.Luo, YuezhenHigh-Resolution Regional Analyses of Daily andHourly Precipitation Climatology over EasternChinaLuo, Yuezhen 1 ; Shan, Quan 1 ; Chen, Liang 1 ; Luo, Yang 1 ; Xie,Pingping 21. CMA Zhejiang Meteorological Bureau, Hangzhou City,China2. NOAA Climate Prediction Center, Camp Springs, MD,USAAnalyses of daily and hourly precipitation climatologyare constructed on a 1kmx1km grid over Zhejiang Provinceof eastern China through interpolating station data withorographic consideration. The high-resolution regionalprecipitation climatology are defined using the PRISMmonthly climatology of Daly et al. (1994), station reports ofdaily and hourly precipitation from 1961 to 2010 at 68stations over the province of ~100K km2, and DEMelevation data. The PRISM monthly climatology used in thisstudy was constructed on a 5kmx5km resolution over Chinawith orographic consideration and represents the meanclimatology for 1961-1990. First, monthly climatology withorographic consideration is defined on a 1kmx1km grid torepresent the mean state of precipitation for 1981-2010. Tothis end, analyzed fields of ratio between monthly stationnormal for 1981-2010 and that for 1961-1990 areconstructed for each calendar month by interpolatingcorresponding station values at the 68 stations using thealgorithm of Shepard (1965). The analyzed ratio is thenmultiplied to the 1961-1990 PRISM climatology to get theanalyzed fields of PRISM monthly climatology for 1981-2010on a 5kmx5km grid. This PRISM monthly climatology isfurther desegregated to a finer grid of 1kmx1km usinglocally established empirical linear relationship betweenprecipitation climatology and elevation. To define analyzedfields of daily climatology, mean daily precipitation is firstcomputed for each calendar day and for each of the 68stations using the station daily reports from 1981 to 2010.Harmonic analysis is performed to the raw time series of365-day mean daily station precipitation and the summationof the first 6 component is defined to represent theevolution of seasonal changes of daily precipitation at eachstation. Analyzed fields of daily precipitation climatology arethen constructed on a grid of 1kmx1km by interpolating thetruncated daily station values and adjusting the interpolateddaily fields against the monthly PRISM climatologydescribed above. The resulting daily precipitationclimatology presents orographic effects passed from thePRISM climatology through the adjustment. To documentthe seasonal evolution of diurnal cycle, analyzed fields ofhourly precipitation are constructed on a 1kmx1km gridover the province for each of the 365 calendar day. To thisend, raw hourly mean precipitation is computed for eachhourly box and for each calendar day using station reportsfor 1981-2010. Reports for a 15-day sliding window centeringat the target day are used in defining the raw hourly meanvalues. The time series of 24 mean hourly values for each94