Variation Of The Atmospheric Water Vapor And Its Radiative Effect Simulations Over The Tibetan Plateau

Based on a variety of different sources of precipitable water vapor (PW) data (ground-basedGPS, radiasonde and numerical weather prediction system analysis), the variations of PW withmultiple time scales and their causes on the Tibetan Plateau (TP) are analyzed using a varietyof different sources of methods. The radiative effects of atmospheric water vapor and otheratmospheric compositions are investigated using radiative models. The results show that (1) theradiosonde (RS) PW is significant dryer than that derived by ground-based GPS (GPS_PW) atLhasa on TP during a period of more than one decade. Different types of radiosonde humiditysensors show different magnitude of the dry bias of PW. The RS_PW dry bias by GZZ-2(goldbeater's skin hygrometer) is more significant than that by GTS-1(carbon hygristor). Thetemporal variation characteristics of the RS_PW dry bias are also investigated. The resultsshow that RS_PW dry bias exhibited pronounced diurnal and annual variations. The solarradiative heating to the humidity sensors may have played an important role in the RS_PW drybias diurnal and annual variability. It can be seen that the diurnal variations of RS_PW dry biasare significant also partly because air temperature is higher at1200UTC than that at0000UTC.The annual variations of RS_PW dry bias are pronounced also partly because air temperature ishigher in summer than that in winter. The calibration methods for the RS_PW dry bias aredeveloped and applied to the GZZ-2and GTS-1sounding PW datasets at Lhasa and Naqu. Thecorrections greatly improve the accuracy of the RS_PW.(2) For long-term changes, the PWand atmospheric mean temperature increased significantly on TP during recent35years (from1976to2010). The increasing trend in PW may be due to the increase in atmospherictemperature during the recent35years. For seasonal changes, PW time series shows variationswith4–14days and60–90days periods during summer monsoon seasons (from early April tolate October). The relationship between PW and sites altitudes can be fitted well using a powerlaw function. For diurnal variation, PW exhibited a pronounced diurnal variation over the TPand its around areas. The characteristics of PW diurnal variation vary with the different sitealtitude, terrain and local climate characteristics.(3) Based on critical phenomena inatmospheric precipitation, the relationship of PW and hourly precipitation is fitted as a powerlaw function. Extreme precipitation probability may increase with the significant increase ofPW and air temperature on TP.(4) The PW comparison between numerical weather model(NWP) system analysis (ECMWF, NCEP, JRA-25, and Met-office) and GPS data reveals that the PW within ECMWF reanalysis data agrees very well with that derived from ground-basedGPS. The PW within JRA-25reanalysis data is slightly underestimated on summer seasons.However, The PW within NCEP reanalysis and NWP output from Met-office have significantdry bias in summer seasons. The effect of the PW differences on surface radiation budget isevaluated using a radiation model. The results show that radiative flux at the surfacedetermined using the model analysis profiles with the water vapor corrected by PWobservations are closer to the observations compared with those without water vapor correction.The radiative flux differences at the surface with and without water vapor correction are largerthan that caused by doubling the concentration of carbon dioxide in the atmosphere in thisregion.(5) The effects of water vapor, ozone, aerosol and cloud on solar radiation exhibitpronounced seasonal variability. The monthly mean solar radiation absorbed by water vapor isabout9~95W/m~2, namely about2%~13%of the global solar radiation. The monthly meansolar radiation absorbed by ozone is about8~12W/m~2, namely about1.51%~1.78%of theglobal solar radiation. The aerosol direct radiative forcing (ADRF) for solar radiation is peak insummer and lowest in winter. The annual mean ADRF is-13.7~-10.1W/m~2. The annual meanCRF is from70W/m~2to140W/m~2and varies with local climate characteristics. The effects ofwater vapor, ozone, and cloud on longwave radiation also exhibit pronounced seasonalvariability. The monthly mean values of the effect is10~78W/m~2, namely6.0%~25.0%of thedownward longwave flux at the surface. The monthly mean value of the effect is1.6～2.0W/m~2, namely about0.6%～1.0%of the downward longwave flux at the surface. The annualmean cloud radiative forcing (CLRF) is from20W/m~2to40W/m~2and varies with localclimate characteristics. Since the seasonal mean PW on summer and the annual mean PWincrease significantly during the period from1976to2010, the seasonal mean global and netsolar radiation on summer decrease with changing rates of-1.16±0.40and-0.93±0.32W m-2/10a respectively. The seasonal mean downward longwave flux on summer increaseswith a changing rate of1.28W m-2/10a. The annual mean global and net solar radiationdecrease with changing rates of-0.57±0.18and-0.45±0.14W m-2/10a respectively. The annualmean downward longwave flux increases with a changing rate of0.65W m-2/10a.