MiKlip first phase: MOSQUITO

MOdel validation uSing high QUalIty verTical profile Observations

Climate model simulations have to be validated by past observations from the real atmosphere. Within MiKlip Module E, MOSQUITO aims to validate MiKlip global and regional hind-casts from MODULEs A to D by comparing simulated and observed profiles of the essential climate variables (ECVs) temperature, relative humidity, wind, and ozone throughout the free atmosphere, from the low troposphere up to the mid-stratosphere. MOSQUITO will use routine meteorological radiosonde and ozonesonde data from a set of German, European and world-wide stations. Contrary to many satellite measurements, profiles from balloon sondes have quite good height resolution. On decadal time-scales, however, radiosonde changes can affect long-term data consistency (Seidel et al., 2004, Free et al. 2005, Haimberger et al., 2008; Steinbrecht et al., 2008, Miloshevich et al., 2009, Sun et al. 2010). In MOSQUITO we will collect radiosonde data, screen for quality, correct for sonde-changes, and use the resulting improved data sets for MiKlip model validation.

Fig. 1: Annual means of temperature from West- and East-German radiosounding stations. No corrections for data inhomogeneities. Since 1990, the same radiosonde type and automatic data recording is used all over Germany.

Time series of radiosonde data allow climate change and variability investigations, e.g. looking at atmospheric circulation patterns, teleconnections, and stratosphere-troposphere connections. These modes are potentially important for long-range weather forecasts and mid-range climate prediction (Thompson and Wallace, 2001; Baldwin and Dunkerton, 2001). In the troposphere, the El-Nino/ Southern Oscillation phenomenon (ENSO) is a major player (Latif, 2006). In the stratosphere, longer-term variations on decadal, inter-annual and weekly to monthly time-scales are well known. They can be attributed to the 11-year solar cycle, to the quasi-biennial oscillation of equatorial winds (QBO), and to vacillations in the strength of the polar winter vortices (van Loon and Labitzke, 1988; Steinbrecht et al., 2006). Radio- and ozonesonde data contain substantial information about these modes of variability, and within MOSQUITO we want to explore this.

Figure 2: Linear trend of temperature as a function of altitude. Blue: for German radiosonde stations from Fig. 1, no correction for possible inhomogeneities. Pink: For world-wide radiosonde data between 30°N and 90°N, partially corrected (see Randel et al., JGR, 2009).


  • Assemble a database of quality controlled profiles of essential climate variables (temperature, relative humidity, wind, ozone) spanning the free troposphere and lower stratosphere, primarily from radiosondes over the last 30 years.
  • Determine and derive important statistical parameters from the database (time dependent mean, median, variance, extremes, probability density).
  • Compare statistical parameters from observations and from MiKlip simulations from Modules A to D.
  • Analyse and compare time series properties of observed and modelled modes of variability (from MiKlip Modules A to D), that are relevant for inter-annual to decadal scale climate predictions (normal modes/ teleconnection patterns, Arctic / North Atlantic Oscillation, ENSO, QBO, solar cycle).

More on MOSQUITONews-Icon


Met. Observatorium Hohenpeißenberg, Deutscher Wetterdienst
Dr. Margit Pattantyús-Ábrahám
Dr. Wolfgang Steinbrecht