Fourth Study Conference on BALTEX Scala Cinema Gudhjem

More documents

Recommendations

Info

- 16 - Vertical Structure and Weather Radar Estimation of Rain Gerhard Peters and Bernd Fischer Meteorological Institute, University Hamburg, peters@miraculix.dkrz.de 1. Introduction Although radar is the most important data source for the aerial distribution of precipitation, quantitative precipitation estimation (QPE) with radar is still far from being satisfying. One major source of uncertainty is related to the fact that the minimum height of the radar scattering volume increases with increasing distance from the radar due to the earth’s curvature. The precipitation observed aloft is usually not identical with the precipitation hitting the surface. The mitigation of this well known ambiguity – often referred to as vertical reflectivity profile (VRP) – is one of the most important steps towards radar based QPE and is the objective of many studies during the last decades (e.g. Koistinen 1991). For the BALTEX area the VRP has been addressed by Michelson and Koistinen (2000) by introducing climatologic range dependent calibration factors on the basis of rain gauge data. A review on radar based QPE with the BALTRAD network can be found in Koistinen and Michelson (2002). The obvious benefit of climatologic calibrations is to eliminate the mean bias. In order to reduce remaining random errors, various approaches are conceivable, which include real-time calibrations, improved knowledge about processes determining the VRP and/or observations of the actual VRP simultaneously with radar reflectivity fields. The two latter approaches would help to establish proxy variables for the VRP, which are observable with radar alone or with auxiliary measurements. Here we report on an on-going study in the frame of the DEKLIM project APOLAS to obtain typical profiles of rain and its micro-physical properties using vertically pointing Doppler radars. We show that the Z-R-relation has a heightdependence, which is particularly distinct at high rain rates (> 20 mmh -1 ), and that its neglect can lead to significant underestimation of rain fall in severe rain events 2. The Data Set Low-power vertically pointing micro rain radars (MRR) operating at 24.1 GHz, designed for long term unattended operation were set up at various sites in and around the Baltic Sea. They provide Doppler spectra of precipitation in range gates of 100 m depth from 100 to 3000 m height with 1 min time resolution. Here we report on results obtained on the peninsula Zingst at the German Baltic coast (54.43°N, 12.67°E). Data spanning an observation period of three years were analyzed. The Doppler spectra were transformed into drop size distributions according to the method described by Atlas et al. (1973). The retrieval algorithm and performance characteristics of the MRR are described in more detail in Peters et al. (2002). In this context a “rain event” is a one minute measuring interval during which rain with more than 0.02 mmh -1 and less than 200 mmh -1 was detected by the MRR. In order to avoid extra difficulties related to the ice phase, only data obtained during the warm season (May to September) within the lower 1500 m height were analyzed. The data were stratified into rain rate classes according to the mean rain rate observed during the corresponding rain event. “Mean” refers to the spatial average within the analyzed height range (100 – 1500 m). Each rain rate class comprises one decade with the limits 0.02, 0.2, 2, 20, 200 mmh -1 . The number of observations falling into each class is given in table 1. Year 2000 2001 2002 Rain Rate mm/h 0.02-0.2 6478 7068 4383 0.2-2 7841 9262 5246 2-20 2422 3742 2354 20-200 96 249 213 Table 1: Number of observations in each rain rate class in each year 3. Data Analysis From the Doppler spectra of each rain event various moments and rain parameters were calculated. The corresponding parameter-profiles were averaged separately for each rain fall class. Results for some parameters are shown in figure 1. In order to get an indication of the representativeness of the results the profiles were averaged separately for each year. The upper row shows the radar reflectivity Z, the basic observed variable of weather radars. (It is derived here from the drop size distribution using the relation Z = ∑i N( D i ) ∆( D i ) D 6 i rather than from the received power, in order to assure comparability with results obtained with longer-wave weather radars, where the Rayleigh approximation of the scattering cross section holds for all drop diameters Di. N(Di)∆(Di) is the number of drops corresponding to the i th line of the Doppler spectrum.) Generally Z decreases slightly with increasing height. The prevailing reasons for this common behavior are probably different in the different rain rate classes. Only a small fraction of the gradient can be explained by the gradient of flux velocity due to the air density gradient. The attenuation is already eliminated in the profiles by applying the single particle extinction coefficient to the drop size distributions and assuming that attenuation is negligible the lowest range gates. At low rain rates the gradient is probably due to the low upper cloud boundary: Drizzle typically originates from shallow clouds in this area. At higher rain rates – at least in the highest rain rate class – the upper cloud boundary is certainly above 1500 m. Here another process must be responsible for the gradient. We believe that the primary reason is here the transformation of the drop size distribution on the fall path. This hypothesis is supported by the mean fall velocity, represented by the first moment of the Doppler spectra in the middle row of figure 1. In the highest rain rate class the mean fall velocity shows a very strong negative gradient although the air density gradient implicates the opposite sign for any given drop size. A possible explanation is that small drops existing aloft are wiped out by coalescence with larger drops, and that this process is not in balance with the break up of large drops. The significance of such process - particularly at high rain rates - is in qualitative agreement with theory (Hu and Srivastava 1994). This transformation of drop size distribution implicates of course a corresponding
height dependence of the Z-R-relation. The third row of figure 1 shows the ratio of rain rates obtained using the dropsize analysis of the MRR divided by the corresponding rain rate, which follows from a fixed Z-R-relation. (Here 1. 42 Z = 350 ⋅ R was used). One recognizes that a fixed Z-Rrelation leads particularly at high rain rates to a height dependent deviation between both estimates. According to these findings high rain rates would be significantly underestimated with weather radar, if Z-R-relations are employed which were established at the surface (as usually). References Atlas, D., Srivastava, R., Sekhon, R., Doppler radar characteristics of precipitation at vertical incidence. Rev. Geophys. Space Phys. 11, 1-35, 1973. Hu, Z., and R.C. Srivastava, Evolution of rain drop size distribution by coalescence, breakup, and evaporation, theory and observations, J. Atmos. Sci., 52, 1761-1783, 1994. 2000 z [m] z [m] z [m] 1400 1200 1000 800 600 400 200 1400 1200 1000 800 600 400 200 0 0 1400 1200 1000 800 600 400 200 0 3 - 17 - Koistinen, J., Operational correction of radar rainfall errors due to vertical reflectivity profile. In: Proc. 25 th Conf. on Radar Met. AMS, 91-94, 1991 Koistinen, J. and D. B. Michelson, BALTEX weather radar-based precipitation products and their accuracies, BOREAL ENVIRONMENT RESEARCH Vol. 7, No. 3: 253-263, 2002 Michelson, Daniel B., J. Koistinen, Gauge-Radar Network Adjustment for the Baltic Sea Experiment, Phys. Chem. Earth (B), Vol. 25, No. 10-12, 915-920, 2000 Peters, G., B. Fischer and T. Andersson, Rain observations with a vertically looking Micro Rain Radar (MRR), BOREAL ENVIRONMENTAL RESEARCH, Vol. 7, No. 4, 353-362, 2002. 5 10 15 20 25 30 35 40 45 50 55 5 10 15 20 25 30 35 40 45 50 55 5 10 15 20 25 30 35 40 45 50 55 Z [dBZ] Z [dBZ] Z [dBZ] 4 5 6 7 v_1.moment [m/s] 8 3 4 5 6 7 v_1.moment [m/s] 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 R/R Z 3 2001 2002 4 5 6 7 8 v_1.moment [m/s] R/R R/R Z Z R [mm/h] 0.02 .. 0.2 0.2 .. 2 2 .. 20 20 .. Figure 1. Mean profiles of various rain parameters, stratified in 4 rain rate classes for 3 years. Upper row: Radar reflectivity. Middle row: Mean fall velocity. Lower row: Rain rate, derived from N(D), divided by rain rate, derived from Z- R-relation. 8
Page 1 and 2: Fourth Stu
Page 3: Preface The 4 th Study</str
Page 7 and 8: - I - Table of Abstracts Title Auth
Page 9 and 10: - III - Sensitivity in Calculation
Page 11 and 12: - V - Parameter Estimation of the S
Page 13 and 14: - VII - The Realism of the ECHAM5.2
Page 15 and 16: Adam, W. K. .......................
Page 17 and 18: - 1 - Activities of the GEWEX Hydro
Page 19 and 20: eference site archive (Cabauw, Neth
Page 21 and 22: - 5 - Remote Sensing of Atmospheric
Page 23 and 24: minute averages around the satellit
Page 25 and 26: - 9 - Precipitation Type Statistics
Page 27 and 28: - 11 - Assimilation of New Land Sur
Page 29 and 30: Multichannel Microwave Radiometer f
Page 31: output has been processed in an equ
Page 35 and 36: - 19 - CERES and Surface Radiation
Page 37 and 38: - 21 - Coastal Wind Mapping from Sa
Page 39 and 40: - 23 - Clouds and Water Vapor over
Page 41 and 42: - 25 - Broadband Cloud Albedo from
Page 43 and 44: - 27 - Observation of Clouds and Wa
Page 45 and 46: shows a ground-track of the vessel
Page 47 and 48: - 31 - Determination and Comparison
Page 49 and 50: - 33 - Spatial Variability of Snow
Page 51 and 52: - 35 - EVA-GRIPS: Regional Evaporat
Page 53 and 54: - 37 - LITFASS-2003 - A Land Surfac
Page 55 and 56: -39- Calibrated Surface Temperature
Page 57 and 58: - 41 - The Marine Boundary Layer -
Page 59 and 60: - 43 - Characteristics of the Atmos
Page 61 and 62: Figure 2. Surface stress from two d
Page 63 and 64: During the experiment the water was
Page 65 and 66: While the number of smallest drops
Page 67 and 68: SCANDIA will not be supported any l
Page 69 and 70: - 53 - Relationships Between Precip
Page 71 and 72: - 55 - Analysis of the Role of Atmo
Page 73 and 74: - 57 - Meteorological Peculiarities
Page 75 and 76: Baltic Sea Inflow Events Jan Piechu
Page 77 and 78: - 61 - The Different Baltic Inflows
Page 79 and 80: - 63 - Observations of Turbulent Ki
Page 81 and 82: - 65 - The Influence of Synoptic Si
Page 83 and 84:
-67- Improved Method for the Determ
Page 85 and 86:
-69- The Helicopter-Borne Turbulenc
Page 87 and 88:
- 71 - CEOP Reference Site Data fro
Page 89 and 90:
- 73 - The BALTEX Hydrological Data
Page 91 and 92:
- 75 - Hydrological and Hydrochemic
Page 93 and 94:
-77- Sea-level Monitoring at MARNET
Page 95 and 96:
1 0.8 0.6 0.4 0.2 GO(CLD-M) ERA40 S
Page 97 and 98:
Figure 2. Monhly mean values of lat
Page 99 and 100:
In the case of an ideal fit, the co
Page 101 and 102:
- 85 - Influence of Atmospheric For
Page 103 and 104:
The largest inter-annual variabilit
Page 105 and 106:
References Andrejev, O, Sokolov, A.
Page 107 and 108:
On the other hand, if the stratific
Page 109 and 110:
Cyberska 1989, Cyberska and Krzymin
Page 111 and 112:
- 95 - Continental-Scale Water-Bala
Page 113 and 114:
- 97 - ICTS (Inter-CSE Transferabil
Page 115 and 116:
The subsurface flow calculation is
Page 117 and 118:
functions only data with higher qua
Page 119 and 120:
- 103 - Modelling the Impact of Ine
Page 121 and 122:
- 105 - Objective Calibration of th
Page 123 and 124:
- 107 - Validation of Boundary Laye
Page 125 and 126:
- 109 - Simulated Dynamical Process
Page 127 and 128:
spreads along isobaths (fig.2). As
Page 129 and 130:
than the precipitation pattern of J
Page 131 and 132:
Figure 2. Long-term variability in
Page 133 and 134:
obvious from temperature charts. Ho
Page 135 and 136:
shortening of the winter seasons by
Page 137 and 138:
Maximum 5 day precipitation (mm) Nu
Page 139 and 140:
scale parameters the correlation be
Page 141 and 142:
similar: the locations of main mini
Page 143 and 144:
- 127 - Storminess on the Western C
Page 145 and 146:
- 129 - Trends in Wind Speed over t
Page 147 and 148:
- 131 - Wind Energy Prognoses for t
Page 149 and 150:
- 133 - Detection of Climate Change
Page 151 and 152:
Figure 3. Cross section of salinity
Page 153 and 154:
of 35 m/s up to 3 hours. Data on wi
Page 155 and 156:
Figure 2. Time series of Mean Sea L
Page 157 and 158:
- 141 - Calculation and Forecast of
Page 159 and 160:
- 143 - An Overview of Long-Term Ti
Page 161 and 162:
- 145 - Baltic Sea Saltwater Inflow
Page 163 and 164:
- 147 - Significance of Feedback in
Page 165 and 166:
Lindström, G., Johansson, B., Pers
Page 167 and 168:
- 151 - Air-Sea Fluxes Including Mo
Page 169 and 170:
- 153 - The Realism of the ECHAM5.2
Page 171 and 172:
- 155 - Figure 3. Accumulated total
Page 173 and 174:
Figure 2. Example for Fourier decom
Page 175 and 176:
Figure1. Modeled percent volume cha
Page 177 and 178:
conditions even more challenging th
Page 179 and 180:
surface heat fluxes close to zero.
Page 181 and 182:
- 165 - Figure 1. Modeled seasonal
Page 183 and 184:
- 167 - Present-Day and Future Prec
Page 185 and 186:
- 169 - Extreme Precipitation on a
Page 187 and 188:
at the stations Pärnu (Estonia) an
Page 189 and 190:
6. Results There are many possible
Page 191 and 192:
and vegetation, and to derive the h
Page 193 and 194:
The structure of water bodies cadas
Page 195 and 196:
Figure 1. The area of Gdańsk subje
Page 197 and 198:
- 181 - Climate and Water Resources
Page 199 and 200:
- 183 - Generating Synthetic Daily
Page 201 and 202:
The evapotranspiration is affected
Page 203 and 204:
Model parameters and coefficients:
Page 205:
3. Hydrodynamic model of nutrient l
Page 208 and 209:
No. 15: Minutes of 8 th Meeting of
show all

Fourth Study Conference on BALTEX Scala Cinema Gudhjem

Create successful ePaper yourself

Delete template?

Save as template?