Persistence and periodicity in the precipitation series of Turkey and associations with 500 hPa geopotential heights

Abstract

Persistence and periodicity in normalised precipitation anomaly series of 91 stations over Turkey were analysed using serial correlation coefficients and power spectra. There were considerable geographical variations and inter-seasonal contrasts with respect to periodicity and persistence characteristics. Lag-one serial correlation (L-1SC ) coefficients for winter series were mostly positive and were significant at the 0.05 level for 31 stations. Annual variations at 17 stations also showed significant positive L-1SC coefficients. In contrast, year-to-year variations in spring series were characterised by negative L-1SC coefficients at most stations (significant at 18). Summer series were characterised by both positive and negative coefficients. Autumn series of most stations were random with regard to serial dependence. For winter series, long cycles of 8.4, 12-12.7, 14, 18 and 21 yr were dominant in the Marmara Transition and the Mediterranean regions, whereas short cycles of 2, 2.1, 3 and 3.2 yr were found for the Black Sea region. Major spectral peaks of most spring series occurred within spectral bands with cycles of around 2, 3, 4 and 5 yr; a 2 yr periodicity was obtained for many stations. Statistically significant negative relationships between precipitation anomalies and 500 hPa geopotential height anomalies in winter and autumn showed an apparent spatial coherence over most of Turkey. Prominent spectral peaks corresponding to about 2 and 3 yr cycles in spring precipitation anomaly series appeared to be associated with similar oscillations in spring geopotential height anomalies. A cycle of 14 yr was found for winter precipitation and geopotential height anomalies.KEY WORDS: Turkey · Precipitation · 500 hPa geopotential height · Persistence · Power spectrum · 'White' and 'red' noise continuum · Periodicity Resale or republication not permitted without written consent of the publisherClim Res 21: 2002 Time-series characteristics of climatological and hydrological observation records can be assigned to the following categories for examination: (1) A stochastic event including random variations or stable processes. (2) Deterministic components corresponding to all the non-random events in the series. Deterministic components include persistence, periodicity, trends, jumps or step-wise changes, catastrophic events (for example, disasters or extreme events), and various combinations of these. (3) A combination of random and non-random events in the series. The third category is important for long-term temporal variations in a series of climatic observations. Based on both real observational records and statistical models, it is well known that trends in climatological series are rarely linear and cycles are rarely purely sinusoidal waves. This peculiarity is found particularly in temporal variations of precipitation series. For instance, a significant downward trend has been revealed for the winter precipitation series at some stations in Turkey (Türkeş 1998). Nevertheless, this trend is not l...

Spatiotemporal variability of precipitation total series over Turkey

“…high-frequency oscillation) that is statistically characterized mostly by a negative autocorrelation (Figure 8). This situation was explained before in detail by Türkeş (1998) and Türkeş et al (2002a).…”

“…A long-term decreasing trend prevails in series of all regions in winter. However, recent years of winter series have been generally characterized by opposite changes towards somewhat increased precipitation (Figure 7), due to low-frequency fluctuations that are dominant long-term temporal characteristics of the 1930 1940 1950 1960 1970 1980 1990 2000 1930194019501960197019801990Years 1930194019501960197019801990Years 1930194019501960197019801990Years 1930194019501960197019801990Years 1930194019501960197019801990 2000 1930 1940 1950 1960 1970 1980 1990 winter precipitation variability in Turkey (Türkeş, 1998;Türkeş et al, 2002a;Türkeş and Erlat, 2005). An evident fluctuation with wet and dry sequences is seen at stations of the MRT, MED, MEDT and CMED rainfall regions that are influenced and controlled more or less by the large-scale pressure and wind systems governing the macro Mediterranean climate throughout the year, particularly its rainfall regime and variability.…”

“…A glance at the Mediterranean regional winter precipitation trends reveals a more detailed picture of the general findings. Some examples of the regional studies dealing with precipitation changes and trends of the Mediterranean region are as follows: Kutiel et al (1996); Brunetti et al (2001); Ramos (2001); Brunet et al (2007);Feidas et al (2007); and for Turkey, Türkeş (1996Türkeş ( , 1998Türkeş ( , 1999Türkeş ( , 2003; Kutiel et al (2001); Türkeş et al (2002a; Erlat (2003, 2005) Türkeş (1996Türkeş ( , 1998, considering the seasonal variability of precipitation and physical geographic characteristics of Turkey (Figure 1). …”

“…In contrast, aridity index values tended to increase significantly towards humid or semi-humid climatic conditions over the northern part of the Continental Central Anatolia region of Turkey. Türkeş et al (2002a) investigated persistence and periodicity components in annual and seasonal precipitation anomaly series at 91 stations of Turkey. Tatli et al (2004) applied a new downscaling method for regional climate process.…”

Precipitation changes and variability in Turkey linked to the North Atlantic oscillation during the period 1930–2000

“…Detailed information for the meta-data and the homogeneity analyses applied to the long-term Turkish precipitation series can be found in Türkeş (1996aTürkeş ( ,b, 1999. The precipitation climatology of Turkey and the long-term variability, trends and changes in the precipitation series were comprehensively investigated previously by Türkeş (1996aTürkeş ( ,b, 1998aTürkeş ( , 1999Türkeş ( , 2003 and Türkeş et al (2002). Copyright  2003Royal Meteorological Society Int.…”