350 likes | 602 Views
ACSYS Final Science Conference 3 St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle - an overview of available data and some results - Bruno Rudolf and Hermann Mächel Global Precipitation Climatology Centre Deutscher Wetterdienst, Offenbach, Germany.
E N D
ACSYSFinal Science Conference3St. Petersburg, 11-14 November 2003The Arctic Hydrological Cycle- an overview of available data and some results -Bruno Rudolf and Hermann MächelGlobal Precipitation Climatology CentreDeutscher Wetterdienst, Offenbach, Germany with contributions from Stefan Hagemann (MPI Met. Hamburg), Reinhard Hagenbrock (Univ. Bonn),Ruediger Gerdes (AWI Bremerhaven) and Thomas Maurer (GRDC Koblenz) with support by the Federal Ministry for Education and Research of Germany (BMBF) within the German Polar and Climate Research Programmes
The Global Hydrological Cycle (fluxes in 1000 km³/yr) is a closed system: global precipitation = global evaporation = 505,000 km³/year of liquid water ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 2
The Arctic Hydrological Cycle Moisture flux Evaporation Precipitation Net ice mass reduction River runoff Net ice mass reduction Oceanic water exchange Moisture flux Surface runoff is not a closed system but has impact from horizontal atmospheric and oceanic water mass exchange ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 3
The Arctic Hydrological Cycle • We need to quantify at least: • River runoff (discharge) • Precipitation (solid, liquid) • Moisture flux convergence • Evolution of the ice mass • Oceanic transports We need models and observations. ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 4
Arctic Runoff Data Base (ARDB) of the Global Runoff Data Centre After establishment of ARDB the number of dailyrunoff data successively increased. One major monthly runoff data set for Russia has been delived in the year 2000 by RHI via Vörösmarty (UNH). ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 5
Arctic Runoff Data Base (ARDB) of the Global Runoff Data Centre • ARDB content for daily data: • June 1993: • up to 40 stations, • period 1977-1988 • June 2003: • up to 240 stations, • period 1900-2000 ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 6
Arctic Runoff Data Base (ARDB) of the Global Runoff Data Centre 270 stations with daily data 2073 stations with monthly data time series end for monthly data: time series end for daily data: ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 7
3787 Arctic Arctic Runoff Data Base (ARDB) of the Global Runoff Data Centre Long Term Mean Annual Surface Freshwater Fluxes into the World Oceans (Updated GRDC Data Product,Version 3) Features: Discharge from catchment through each 1/2° coastline grid cell (11.853 cells) Based on 251 GRDC stations >25.000 km2 + 1378 smaller stations Extrapolation via a simple runoff coefficient estimation (using GPCC precipitation data) Discharge from arbitrary coastline sections by simple aggregation GRDC estimate of total runoff from the Arctic catchment area: 3787 km³/year More about ARDB is given by Poster 68 of T. Maurer (GRDC). ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 8
Arctic Precipitation Data Archive (APDA) What have we got until 2003 ? 1) Increased number of stations with precipitation data Region _______ FSU/Rus. Canada Norway Alaska GPCC status 1994resolution period stations monthly 1986-1992 622 monthly 1986-1992 320 monthly 1986-1992 77 monthly 1986-1992 153 APDA status 2003 stations resolution period 2004 daily 1891-1999 7281 daily 1840-2000 649 monthly 1950-2000 514 daily 1891-2000 2) In addition, APDA collects snow data You will find more about APDA on the posters 72 and 73 of Mächel & Rudolf 3) Results of the WMO Solid Precipitation Measurement Comparison Study (Goodison, Louie & Yang 1998) 4) Full global satellite-based precipitation estimates ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 9
Systematic gauge measuring error Precipitation measured by gauges is systematically underestimated because of evaporation, wetting losses and drift of snow and drops by wind across the gauge funnel. In order to get reliable global or regional precipitation amounts, an adequate correction of the data used or of the product is required. (Fig. after SEVRUK 1989) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 10
Systematic gauge measuring error WMO Instruments Comparison Programme Double Fence International Reference Here: Comparison Site at Barrow/Alaska, June 2002 ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 11
Systematic gauge measuring error WMO Solid Precipitation Measurement Comparison Study Catch ratio in % of the DFIR as a function of wind speed for various gauge types, here for dry snow The undercatch is characterized by the instrument type. (Fig. from Goodison et al. 1998) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 12
Systematic gauge measuring error WMO Solid Precipitation Measurement Comparison Study Precipitation phase and wind speed are the most important meteorological parameters for the systematic error The GPCC has developed a method to estimate wind speed, precipitation phase, air temperature and humidity from synoptic data, which are needed to calculate the bias corrections on a daily “on event“ basis. (Figure: T. Günther in Goodison et al, 1998) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 13
Systematic gauge measuring error A new bias correction method for global data Mean percentual correction for all SYNOP precipitation based on the new „on event“ correction method Amazon catchment: mean bias = 4% BALTEX catchment: mean bias = 25% Comparison of monthly percentual corrections in % of observed data derived from daily corrections for the years 1996 and 1997 and long-term mean monthly corrections after Legates ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 14
Comparison studies • Comparison of gridded precipitation data • based on different gauge data collections: • GPCC Monitoring (realtime data) • GPCC Full Data (including additional data) • CRU Climate Research Unit Norwich/UK • FAO UN Food and Agricultural Organization • GHCN Global Historical Climatology Network • FSU/Russia (data prepared by Groisman) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 15
Comparison of monthly precipitationJanuary 1987 (80-90°E, 50-60°N) FAO GPCC full data FSU 60°N 55°N 50°N 60°N Number and distribution of stations not available 55°N GHCN GPCC monitoring CRU 80°E 90°E 80°E 90°E
Comparison studies Data coverage for the box 80-90°E, 50-60°N 1987 ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 17
Comparison studies River catchment area-mean precipitation versus river runoff observations ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 18
Comparison of annual precipitation and runoff (river Lena) annual ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 19
Comparison of monthly precipitation and runoff (river Jenisei) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 20
Comparison of satellite-based precipitation estimates Profiles of zonal mean precipitation (Dec.1989 – Feb.1990) Land surface Outliers are IR-GPI & SSMI-scat, but adjustment to in situ is possible. Improvement by calibrated TOVS. Global ocean Outliers are MSU in general, SSMI-scat in mid-high latitude, and tropical ERA-40 ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 21
The Arctic Hydrological Cycle in numerical models Model sources until 1994: Output from operational global weather forecast models has been used to estimate unmeasured variables. The models generally showed extreme large spin-up effect in precipitation forecasts. Developments during the period 1994 to 2003: NWP Model Re-analyses (NCEP-1, ERA-15, NCEP-2, ERA-40) provided 4dim gridded meteorological data sets based on available meteorological observations interpolated and balanced by the model data assimilation and forecast. But how well do they describe the hydrological cycle? ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 22
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Comparison of ERA-40 precipitation forecasts and observations (GPCP V2 Sat-Gauge) Period 1991-1995 In this period, ERA-40 overestimates tropical precipitation, in particular over ocean. ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 23
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Differences of precipitation ERA-40 minus GPCP V2 Sat-Gauge January 1995 July 1995 mm/mon 60° 60° 50° 50° ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 24
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Differences of zonal area-mean precipitation for ERA-40 minus GPCP V2 Sat-Gauge ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 25
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 The ERA-40 hydrological cycle over the Arctic Ocean catchmentStefan Hagemann, MPI-Met. Hamburg With regard to the observational data used, the ERA-40 period consists of three phases: 1958-1972: the pre-satellite phase 1973-1988: the transitional phase 1989-2001: the satellite phase ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 26
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Mean annual cycle of monthly precipitation for 8 Arctic rivers: Mackenzie, N.Dvina, Pechora, Indigirka, Lena, Jenisei, Ob, Kolyma Mackenzie: Large differences occur for ERA-40 strings 1958-1972 and 1973-1988 in summer For all figures: The difference between GPCC (uncorrected) and GPCP come from the application of gauge bias corrections after Legates applied in GPCP Indigirka: Differences ERA and Obs are large because of the smaller size of the river basin Stefan Hagemann MPI Met. Hamburg ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 27
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Comparison of catchment area mean precipitation for 8 Arctic rivers, the area of the 6 largest, and the full Arctic catchment for ERA-40 and Observations (GPCC p3, CRU p1-2) Full Arctic catchment 6 largest rivers This study should be repeated based on the ACSYS data base. Stefan Hagemann MPI Met. Hamburg ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 28
The Arctic Hydrological Cycle of NWPM Reanalysis ERA-40 Comparison of river runoff for 8 Arctic rivers and the area of the 6 largest for ERA-40 and Observations (GRDC) 6 largest rivers Stefan Hagemann MPI Met. Hamburg ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 29
University of Bonn Moisture flux convergence Horizontal distribution of vertically integratedmoisture flux convergence (= P-E) Shown here: the average 1979-1993 based on mass consistent radiosonde data, smoothed to T42 (Hagenbrock 2003) Posters on this topic are presented e.g. by Hagenbrock & Hense (47) and Serreze et al. (85) mm/day ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 30
University of Bonn Moisture flux convergence Vertically integratedmoisture flux convergence, average 70°-90°N Comparison of ERA-15 and radiosonde P-E average 1979-1993 mean annual cycle time-series 1979-1993 radiosonde data: average: 0.45 mm/d ERA-15 reanalysis data: average: 0.48 mm/d Cullather et al.: radiosonde data: average: 0.45 mm/d Cullather et al.:ERA-15 reanalysis data: average: 0.50 mm/d (Hagenbrock 2003, Univ. Bonn) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 31
Less precipitation or runoff ? No, precipitation increase observed. Less sea ice melting ? No, ice melting increased, too. Import of salt through the straits ? Yes, the higher NAO index has led to stronger transports of salt water from the Atlantic into the Arctic Ocean. Arctic Ocean Freshwater Balance A decrease of freshwater content has been observed from 1965 to now.Why ? Project ArcticFW: Freshwater Balance of the Arctic Ocean: Long-term Variability and possible future development Ruediger Gerdes et al., AWI Bremerhaven km³ Increase of the Arctic Ocean Salinity: ~ 0.07 ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 32
The Arctic Hydrological Cycle Some Arctic hydrological budgetcomponents from the ERA-40 verification study for the full Arctic hydrological catchment area (land surface only) ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 33
The Arctic Hydrological Cycle Moisture flux Evaporation Precipitation Net ice mass reduction River runoff Net ice mass reduction Oceanic water exchange Moisture flux Surface runoff Estimates of all budget components and fluxes shall be compiled and discussed based on the output of this conference. Call for average budget data in order to fill in the diagramme ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 34
The Arctic Hydrological Cycle • Looking forward • Research on Arctic Climate must continue beyond ACSYS. • Hydrometeorological variables (precip, snow, evap, runoff) • are still difficult to quantify. Future topics of CliC concerning hydrometeorological observation might be: • Data collection and analysis: • Updates and additional historical data are necessary. • New automatic precipitation gauges: • Meta data and new comparison studies are required. • Satellite-based observation: • The Global Precipitation Mission (GPM) will supply new • observations for high latitudes. Verification is needed. • NWPM Reanalyses: • ERA-40 should get a corrected update. ACSYS Final Science Conference St. Petersburg, 11-14 November 2003 The Arctic Hydrological Cycle Bruno Rudolf, Hermann Mächel 35