Türkiye'de ekstrem sıcaklık trendleri
Extreme temperature trends over Turkey
- Tez No: 66516
- Danışmanlar: DOÇ. DR. MİKDAT KADIOĞLU
- Tez Türü: Yüksek Lisans
- Konular: Meteoroloji, Meteorology
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1997
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Meteoroloji Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 110
Özet
ÖZET Yaşantımız boyunca anormal hava olaylarıyla karşılaşıldığında akla ilk olarak 'iklim değişiyor mu?' sorusu gelmektedir. Günümüze kadar yapılan birçok çalışmalarda, gözlenen meteorolojik anomalilerin iklim değişimine bağlanmış olduğunu görmekteyiz. Bu çalışmada, maksimum ve minimum sıcaklık ekstrem değerlerin sayılarındaki değişimi tespit etmek amacıyla verilerdeki hatalara daha az duyarlı olan rekor analizi ele alınmıştır. Türkiye geneline dağılmış 49 DMİ meteoroloji istasyonun ortalama 60 yıllık periyod boyunca gözlenen aylık ortalama maksimum ve minimum sıcaklık verileri kullanılmıştır. Bu analizde öncelikle, her istasyon için belirli bir zaman periyodunda ekstrem sıcaklıklarda istatistiksel olarak kırılması beklenen rekor sayılan belirlenmiştir. Her istasyonda gözlenen rekor sayılarının bu beklenen rekor sayısından olan farkları hesaplanmıştır. Bu farkların istatistiksel olarak önem seviyesini belirlemek için frekans histogram diyagramları analiz edilmiştir. Elde edilen diyagramlara uygun normal dağılım diyagramı geçirilerek %5 istatistiksel önem seviyesine göre değişimin kritik değeri belirlenmiştir. -2 ile 2 arasındaki değerler anlamlı, bunların dışında kalan değerler de istatistiksel olarak anlamsızlığı belirtmiştir. Bölgeler üzerinde değişimin nerelerde olduğunu görmek için beklenen ile gözlenen rekor sayılarının farklarını gösteren haritalar analiz edilmiştir. Bu haritalara göre, belirgin bölgesel değişimlerin fazla olmadığı, zamansal olarak mevsimsel değişimlerin daha belirgin olduğu görülmüştür. Ekstrem minimum (gece) sıcaklıklar, ekstrem maksimum (gündüz) sıcaklıklarından daha fazla değişim göstermiştir. İlkbahar ve kış mevsiminde gece net olarak ekstrem yüksek sıcaklıkların sayısında (rekor sayısı) artış olduğu belirlenmiştir. Sonbahar mevsiminde ekstrem gece sıcaklıklarında beklenen yakın rekor kırma sayılan elde edilmiştir. Yaz mevsiminde ise özellikle Karadeniz'de gece %5 önem seviyesinde ekstrem sıcaklıklarda azalma görülmüştür. Bu şekilde yapılan incelemede, genel olarak istatistiksel anlamda %5 önem seviyesine ulaşmış rekor farklan, ekstrem sıcaklıkların değişimi hakkında yorum yapılmasını sağlamıştır. Genel olarak ekstrem sıcakhklardaki değişimler ile ortalama sıcakhklardaki değişimler uyum içindedir. Fakat bazı durumlarda uyumsuzluk görülmüştür. Bunun için rekor sayılarının, teorik olarak hesaplanan beklenen rekor sayışma göre değişimi yerine kendi zaman serisi içindeki değişimlerine bakılmıştır. Maksimum ve minimum sıcaklıkların başlangıç değerinin üstünde (yüksek) ve allında (düşük) kalan değerlerinde ardışık 10 yıllık aralıklarla kaydırarak hesaplanan ortalama sıcaklıklar ile aynı periyod boyunca gözlenen rekor sayılannın dağılımı incelenmiştir. Buna göre ekstrem sıcaklık rekorlarının her zaman önemli sıcaklık değişimine işaret etmediği görülmüştür. Örneğin rekor sayılan, ekstrem sıcaklıklara göre azalma trendi gösterse de, ortalama sıcaklıklara zaman serisinde artış eğiliminde olabildiği gözlenmiştir. XI
Özet (Çeviri)
SUMMARY EXTREME TEMPERATURE TRENDS OVER TURKEY Most of the studies in the meteorology and climatology are used to be based on the idea that 'climate is stationary' up to the mid 1950s. The climate normals were also not allowed to change in a long time. However, glacial period is the most important proof of the climate change. Nowadays, because of climate system has natural variability, climatologists agree with the idea that climate is never stationary. Besides natural variability, human activities have entered a new phase that climate is effected in regional and global scales. Emissions are caused by several human activities, are speeded up increasing greenhouse gases. One component of the 'popular vision' is the notion that the high temperatures of the summer season will increase in magnitude and in frequency. According to speculations, greenhouse warming will add heat to all air masses at all times of the day, and therefore, there will be substantial increases in the frequency of record- breaking high temperatures (Balling, 1992). This prediction is not just an exaggeration of the global warming issue; it is well founded in the professional literature. In particular, published work by Mearns et al. (1984) and Hansen et al. (1988) leads directly to the conclusion that global warming will increase the frequency of extreme high temperatures. This issue becomes especially important and threatening given the linkage between extreme high temperatures and human health, agricultural productivity, and water and energy demand. In another research project on the maximum temperature issue, a different approach to test the hypothesis is suggested that general warming will increase the frequency of extreme maximum temperatures. Balling et al. (1992) examined the relationship between the summer mean temperature in Phoenix and frequency of extreme maximum temperatures in the city. Their results showed the rapid warming in Phoenix increasing the occurrence of moderately high maximum temperatures, but somewhat surprisingly, the rapid warming was having only a small effect on the occurrence of the extreme maximum temperatures. The rather simplistic scheme used by Hansen et al. (1988) for predicting massive increases in very high temperatures in United States cities would seriously overestimate the occurrence of the extreme maximum temperatures in Phoenix. Balling et al. (1992) concluded that 'caution (should) be exercised in predicting distributions for extreme maximums and minimums from projected increases in mean monthly or seasonal temperatures '. As a result of warming the atmosphere in global scale, recently several investigations on how is climate change and how is societies effected from climate xuchange are examined. Various trend-detection studies have been carried out in different parts of the world, mostly for identification of climate change. Some of these cases have shown significant trend components, especially during the last 40-year period (Karl et al., 1993). In general, statistical parametric methods such as the moving average (Jones and Jiusto, 1980) and linear regression methods, as well as two-phase regression models (Solow, 1987) are used in previous studies. Parametric and non-parametric tests, are used for testing whether there have been statistically significant trends. Parametric methods require the assumption of a normal distribution. However, in a non-parametric test, such as the Mann-Kendall rank statistic, the departure from a Gaussian normal frequency distribution is not a concern (Mitchell et al., 1966). However, the physical interpretations have been related, at times, to the greenhouse effect, global warming, urban heat islands and to aerosols that exert cooling effects on our climate (Balling, 1992). The effects of urbanisation on temperatures and precipitation also have been discussed widely in the literature (Chagnon, 1992). Therefore record analysis that is non-parametric method and only is concerned with the extreme values is used. Nowadays, flash floods, unusual weather patterns, abnormal high temperatures are caused same questions like ' has the weather gone mad? ' and ' have the greenhouse effect and ozone getting hole changed our climate? '. Of course, these questions are based on strong assumptions. Observed meteorological parameters are smaller or larger than observed parameters in previous years and can reach record value. In most studies, the meteorological records are related to climate change. They can be, however, also related to the nature of weather for very random fluctuations to occur, and there is a record broken weather event every day somewhere in the world. Extreme temperatures instead of mean temperatures cause enormous social effects. It can be questioned that whether or not the extreme temperatures are related to climate change? Do record values cause significant temperature change? In order to understand this, in this study, the only record-breaking minimum and maximum temperatures were analysed. The extreme high and low temperature studies, use the highest and lowest observed temperatures (extremes) within each month compiled for stations within the country. This makes this thesis work different from the classic extreme temperature analyse studies. The record-breaking values are also less sensitive to classic time series problems such as changes of gauge-site, height, exposures of instrument shelter and urbanisation. It was also investigated whether the observed record values caused significant mean temperature change over Turkey. In this thesis forty-nine stations were chosen for the study of extreme temperature trends over Turkey. The data consist of the observed monthly mean maximum and minimum temperatures recorded in these stations. None of these stations had more than 5 per cent of its data replaced with estimates from nearby stations because of missing observations. These stations are located mostly in small towns, but some stations are within or near to large urban areas, which may be influenced by the urban heat island. For instance, Göztepe meteorology station in Istanbul falls within the suburban area and therefore it is influenced by the heat island xuieffect due to the high population concentration. It is well known that the urban heat island often tends to manifest itself strongest during night-time hours (Landsberg, 1981). While investigations are done, it is important that data must be reliable. Homogeneity tests were done to the this reliability of this data by many authors. The Swed and Eisenhart (1943) run test is a non-parametric procedure used in determining the homogeneity of a time series. It can adequately detect, in general, inhomogeneities as well as other problems, such as the change of instrument, relocation of stations, etc. However, after detection the inhomogeneity, in order to identify what is the major cause, further detailed physical and/or meteorological studies are needed. Vining and Griffiths (1985) have used this technique to study the climatic variability at 10 stations in USA. The data used in this study were tested by various homogeneity tests by Türkeş et al. (1995), Toros (1993), Özçelik (1996) and Kadıoğlu (1993,1997). While we analyses, no matter what the present temperature record may be, it is certain that eventually there will occur a wanner year. No matter how many record breaking years have been counted to date, there will be always one more. Such unit increments make the count of record breaking years arbitrarily large as the years of observation increase indefinitely. For example, where I live: what will be a new maximum probability in the next year? Temperatures are less likely later than early in a sequence of observations. The first observation necessarily must be a 'record high'. But, prior to observing any values, that is known that the second of two numbers in random sequence has equal probability of being smaller or larger than the first one. Hence, the probability is exactly 1/2 that a second, independent observation will be a new record high surpassing the initial record, assuming that there cannot be an exact tie (if measurement is arbitrarily precise). From the same perspective, there is probability 1/3 that a third trial will be a new maximum, since the last of three repeated observations is equally likely to be smallest, middle, or largest. Similarly, all 10 ranks are equally likely for the tenth observation; so maximum rank for the tenth observation has probability equal to 1/10. The theoretical expected or average number of record highs in a chronological sequence of n-independent observations is the sum of these probabilities: 1 + 1/2 + 1/3 +... + 1/n Record breaking annual event depends on the period of record or observation. Consider a case in which the records just begin with an observational program at some station and then record breaking cold winters, will have the expected value as 1 + 1/2 +1/3 +... + 1/10 = 2.92. XIVAssuming a conventional 70-year lifespan, it is necessary to continue the sum of successive fractions up to 1/70. These fractions are the reciprocals of the whole numbers from 1 to 70 and the sum turns out to be nearly 5. 1 + 1/2+ 1/3 +... + 1/70 = 5 If an average life-expectancy were 100 years, the sum of reciprocals would still be only a little more than 5. 1 + 1/2 +1/3 +... + 1/100 = 5 In this study, high and low record numbers of mean monthly maximum and minimum temperatures are determined by using the record analysis method. Differences between expected and observed record numbers are examined. Therefore the expected and observed record numbers are determined at each of the 49 stations for every month and season. Significance level of the differences between expected and observed records are also determined statistically. By using standard normal distribution and corresponding to the 5 per cent significance level confidence the limits is determined as ±2, which shows significant areas of trends over maps. The record differences between expected and observed records in maximum and minimum temperatures are summarised in Tables Al and A2, respectively. These tables include station locations according to sequence based on the degrees latitude from north to south. Herein, seasons are defined as 3-month periods, namely spring (March- May); summer (June-August); autumn (September-November) and winter (December-February). After critical value is determined as %5 statistically contour maps are drawn that show record differences which reach to significance level. According to the differences that have reached to %5 significance level, extreme temperature changes are examined. After that, observed record values of maximum and minimum temperatures, are added as successive cumulatives. Later maps, showning differences among cumulative sums are drawn. According to these cumulative sum maps, it is said that if these differences are positive, then there is warming trends in mean temperatures; if these differs are negative, there is cooling trends in mean temperatures. The trend in extreme temperature data do not show distinctive meridional patterns across Turkey, but temporal patterns are relatively more obvious. The non appearance of meridional trends is due to heterogeneous topographic effects as well as surface features such as forest, different types of rock outcrops, different sizes of urbanisation centres and climate regions. XVA close inspection of the results on extreme minimum temperatures indicates that almost all the stations have warming trends at night during spring and winter seasons. It is clear from Table A.2 that during spring and winter months, the linear trends are statistically significant at the 5 per cent levels. The changes in extreme maximum temperatures are less than extreme minimum temperatures. The relationship between changes in mean temperature and associated changes in the probabilities of extreme temperature events is inherently non-linear, reliance on mean temperature changes essentially precluding any consideration of extreme temperature events when performing climate impact analysis. In view of the importance of extreme events, a useful exercise in climate impact analysis would be to analyse the characteristics of daily temperature time series to determine possible ranges of changes in probabilities of certain extreme temperature events associated with changes in the mean of the series. Our results are intended to illustrate, by concrete examples, the possibility that potential long-term changes in mean temperature may exert their principal impacts on the environment and society through changing probabilities of extreme temperature events. Although the scope of the thesis work is limited to consideration of only the likelihood, not the impacts of certain climate events, the importance of dealing explicitly with extreme events in climate impact studies has been demonstrated by Mearns et al. (1984). Similarly, in this thesis, according to results, trends of the mean maximum and minimum temperatures agree with trends of the extreme maximum and minimum temperatures, but, at some stations, while mean temperatures show warming, extreme temperatures indicate cooling. XVI
Benzer Tezler
- Türkiye'nin kuzey ve güney kesiminde yer alan şehir/kır yerleşmelerinde ekstrem sıcaklık indislerinin değişimi
Change of extreme temperature indices in urban/rural settlements in the northern and southern parts of Turkey
DİLEK AYKIR
- Yüzey, 850, 700 ve 500 hpa basınç seviyelerinde sıcaklık ile nem parametrelerinin karşılaştırmalı trend analizi: Türkiye örneği
Comparative trend analysis of temperature and humidity parameters at surface, 850, 700 and 500 hpa pressure levels: Case of Turkey
SIRACETTİN GÖZALAN
- Türkiye'nin Doğu Akdeniz kıyılarında sıcaklık ve yağış trend analizleri ve ekstrem hadiseler
Analys of temperature and precipitation trends and anomalous phenomena in the Eastern Mediterranean coasts of Turkey
GÜLTEN İÇEL
- Yüksek çözünürlükteki bölgesel iklim modeli simülasyonu ile ekstrem iklim indekslerinin belirlenmesi
Defining extreme indices by using high resolution regional climate model simulation
FULDEN BATIBENİZ
Yüksek Lisans
Türkçe
2014
Meteorolojiİstanbul Teknik ÜniversitesiMeteoroloji Mühendisliği Ana Bilim Dalı
DOÇ. DR. BARIŞ ÖNOL
- Urbanization's role in shaping temperature records: Insights from İzmir, Türkiye
Sıcaklık rekorlarının şekillenmesinde kentleşmenin rolü: Türkiye, İzmir'den çıkarımlar
FATMA BAŞAK SAKA
Yüksek Lisans
İngilizce
2024
Meteorolojiİstanbul Teknik ÜniversitesiMeteoroloji Mühendisliği Ana Bilim Dalı
PROF. DR. YURDANUR ÜNAL