Vol. 119, No. 1 * Pages 1–128 * Januar - March 2015

Three different regional climate models (DMI-HIRHAM5, HadRM3, and KNMI-RACMO2) driven by ERA-40 reanalysis and also driven by global climate models (GCMs) obtained from the EU-Ensembles project have been compared to observed data over the Iberian Peninsula (IP) to assess the accuracy of simulated precipitation and temperature. KNMI-RACMO2 and DMI-HIRHAM5 were the best models for accurately simulating precipitation and temperature, respectively, although large uncertainties still affect their simulations. The same RCM simulations driven by GCMs have been used to project the seasonal expected changes in precipitation and temperature for the periods 2011–2050 and 2051–2090 relative to 1961–2000 under the A1B climate change scenario. From the results, a clear decrease in mean precipitation is expected in most IP for spring, summer, and autumn, but no clear signal was found in winter. Moreover, future projections showed a large increase in mean temperatures in all seasons being more evident in the interior of the IP especially in summer. The decrease in mean precipitation and the increase in mean temperature projected for the IP, could worsen current drought conditions especially for the second half of the 21st century.

Nitric oxide soil emission flux was measured by 2–2 parallel manual and auto dynamic chambers on hourly basis above a Hungarian semi-arid, sandy grassland between August 2012 and January 2014. The measured datasets covered 43–85% of time period depending on chambers. We applied a gap filling method based on multivariable analysis (Sigma Plot) combined with maximum likelihood method. Trend of gap filled dataset shows large peaks mostly in summer and early fall. When soil parameters are far from the optimum (dry, warm conditions), the fluxes are negligible. Application of manual chambers closed for longer period results in substantial positive bias in flux estimation compared to auto chambers as a consequence of measurement setup, different temperature, and drier soil conditions below the chamber. Mean fluxes applying permanently closed dynamic chambers are approximately three times higher compared to auto chambers: 0.176±0.489 nmol m^–2s^–1 and 0.058±0.130 nmol m^–2s^–1, respectively.

This study describes a weather-radar-based investigation of long-lived thunderstorms in Hungary in the period of 2004–2012. An objective method was developed for identifying and tracking convective cells. The cells were represented by so-called thunderstorm ellipses. In this research, intensive objects were classified into 3 categories such as severe, highly severe, and extremely severe thunderstorm ellipses. The categories were defined by radar reflectivity thresholds 45 dBZ, 50 dBZ, and 55 dBZ. Only those cells were involved in the investigation whose lifetime extended more than 1 hour. In the 9-year period, 2625 severe, 597 highly severe, and 45 extremely severe long-lived thunderstorm ellipses were found. Stronger cells moved faster and at most intensive cells, right-turning movement was more frequent. Many of these long-lived, strong objects could be supercells. The applied methods and results can be used for severe weather forecast and nowcasting in the Carpathian Basin.

Spectral analysis is often based on a comparison of the periodogram and the spectral density of a so-called background noise. This spectral density is estimated by fitting a first order autoregressive (AR(1)) process to data, as climatological time series generally exhibit red noise spectra that can be approximated by AR(1) models. When periodogram exceeds some threshold at a frequency, the spectrum is said to differ from this background noise, and the frequency is characteristic for the time series in question. The traditional periodogram, however, must not be used without modifications for unevenly spaced data. Additionally, red noise, characterized by spectral densities monotone increasing to low frequencies, covers a much wider class of processes than the AR(1) processes. Our purpose is (1) to introduce a new periodogram (ELSP) based on a least square (LS) fit for an entire set of frequencies instead of using the well-known Lomb-Scargle periodogram (LSP) based on individual LS fits for individual frequencies; (2) to estimate the spectral density without any assumption on its analytical form using the nearly isotonic regression (NIR) method with either ELSP or LSP. As NIR allows the possibility of deviations from red noise, comparison of the periodogram with a background noise is unnecessary. Note that ELSP has never been used before as is a new concept for defining the periodogram for unevenly spaced data. NIR is more or less known for curve fitting problems but has not been applied yet to spectral density estimation. Three examples show that although ELSP does not radically differ from LSP, NIR-ELSP and NIR-LS spectra can exhibit distinct shapes.

The cycle properties of the annual average, absolute maximum, and absolute minimum precipitation values have been calculated from precipitation data the Mátra and Bükk regions. The cycle parameters of annual average and annual absolute maximum precipitation values have been determined using the data of a shorter 34-year (1970–2006) and a longer 53-year (1960–2012) period (38 precipitation measurement stations) through the determination of the parameters of frequency, amplitude, and phase with an analytic version of the discrete Fourier transform (DFT), and the values obtained on the basis of the two periods have been compared. Using prognosis parameters, a prognosis until 2025 has been made. Then, the regression function of the variation in time of average and absolute maximum precipitation values has been determined on the basis of actual and prognosticated data for the whole period (1960–2025).

The aim of the study was to estimate long term evapotranspiration (ET) of Kis-Balaton wetland through the investigation of Lake Fenéki. Data set was processed using the West-transdanubian Water Inspectorate methodology. Potential evapo-transpiration (PET) was calculated using Hungarian empirical models (Antal and Dunay), while Lake Fertő formula was applied evaluating the ET that includes the impacts of vegetation.
Calculated PET values of the wider (adjacent) environment of Lake Fenéki (Zalaegerszeg, Nagykanizsa, and Keszthely meteorological stations) differed significantly and further variation was observed in PET, when measured meteorological elements on Lake Fenéki were applied. PET increment, as a result of linear trend fitted to the 43-year long data (Keszthely station) was 3–4 mm year^-1. Relation between PET calculated from the data of Keszthely station and for Lake Fenéki was strong, so PET of Lake Fenéki can be originated from the data of Keszthely station.  Calculated ET was not significantly different due to the likely similar input data in ET calculation model of Hungarian Meteorological Service (OMSZ).
43-year annual mean ET for Lake Fenéki was 809±88 mm. This ET was 84% of calculated PET. Analyzing the nine dry-warm seasons, average annual ET exceeded the long term average (874.7±37.6 mm) with 78 mm. The average ET of the remaining 34 wet-cold seasons totalled 796.6±89.4 mm.
Empirical formulas cannot be replaced, according to monthly ET comparisons, by using “A” class pan estimating the ET of aquatic habitats.
Seasonal pattern of monthly ET time series for Lake Fenéki was analyzed using autoregressive integrated moving averages (ARIMA) modeling technique. After first differencing, the transformed series was stationary and found to be governed by moving average process of order 1.

In our work, we planned deeper understanding of cold drops, closed air masses separated from the main western stream, by using ECMWF ERA-Interim reanalysis and ensemble forecasts. Upper level low (ULL) recognition algorithms were used to study 70 independent cold drop occurrences from the last decade in the middle and eastern European region. This led to the ascertainment of the usual location of ULLs in relation to Hungary, their core temperature, axis lean, horizontal temperature change, and identification on „plum” diagrams. Our studies included the usage of the potential temperature of the 2PVU (2 potential vorticity unit) surface, potential vorticity field related to 315 K, and 300 hPa wind speed These recommended new variables are available from operational deterministic and ensemble forecasts, and their usage is highly effective, hence making the identification of cold drops a lot easier than before.

Jean-François Geleyn passed away on  8 January, 2015. We will always remember him as a scientist and friend who was always supporting us and our meteorological community. We will miss Jean-François with sorrow, but we will preserve his memory and put forward his legacy in Numerical Weather Prediction.