Geri Dön

Klimatolojik serilerden Türkiye ikliminde trend analizi

Başlık çevirisi mevcut değil.

PDF İndir

  1. Tez No: 39425
  2. Yazar: HÜSEYİN TOROS
  3. Danışmanlar: YRD. DOÇ. DR. MİKDAT KADIOĞLU
  4. Tez Türü: Yüksek Lisans
  5. Konular: Meteoroloji, Meteorology
  6. Anahtar Kelimeler: Belirtilmemiş.
  7. Yıl: 1993
  8. Dil: Türkçe
  9. Üniversite: İstanbul Teknik Üniversitesi
  10. Enstitü: Fen Bilimleri Enstitüsü
  11. Ana Bilim Dalı: Belirtilmemiş.
  12. Bilim Dalı: Belirtilmemiş.
  13. Sayfa Sayısı: 191

Özet

Günümüzde artan aşırı şehirleşme, özellikle sanayi ve yerleşim bölgelerinden çıkan gaz atıkları, motorlu vasıtalar. ısıtmada kullanılan yakıtlar ile bitki örtüsünün değişmesi, bir bölgedeki ısı dengesini ve hidrolojik çevrimi etkileyebilmektedir. Bir bölgedeki sıcaklık ve yağıştaki salınım ve değişim trendlerinin bilinmesi; yapılaşmadan, kurulacak endüstriye kadar bina ve aletlerin dizayn ve yapımında önem kazanmaktadır. Yine iklimdeki değişim ve salınımların bilinmesi, tarım ve orman ürünlerinin artımı açısından değerli faydalar sağlayacaktır. Dünyada iklim değişimi ile ilgili çalışmalar şu ana kadar daha çok sıcaklık ve yağış verileri incelenerek yapılmaktadır. Küresel ölçekte, bilhassa sıcaklığa bağlı olarak meydana gelen iklimsel değişimlerin bölgesel olarak incelenmesinde faydalar vardır. Bu çalışmada, bölgesel olarak iklimde bir değişimin olup olmadığını belirleyebilmek için Türkiye genelini temsilen 18 meteoroloji istasyonuna ait yağış ve sıcaklık verileri incelenmiştir. İncelenen iklim verilerinin kalite kontrolü için tüm veriler homojenlik testlerine tabi tutulmuştur. Verilerin normal dağılıma uyup uymadıkları gibi bazı istatistiki özellikleri de tesbit edilmiştir. Çalışmanın esasım teşkil eden iklim değişimi için en küçük kareler metodu yanında dağılımdan bağımsız olan Spearman mertebe korelasyonu ve Mann-Kendall mertebe korelasyonu test istatistiği uygulanmıştır. Yapılan homojenlik testlerine göre incelediğimiz veriler genelde homojendir. Trend analizlerinin sonuçlarına göre, Türkiye genelinde incelenen istasyonlarda yüksek sıcaklık ve yağışta genelde belirgin bir trend gözükmezken, özellikle ilkbahar düşük sıcaklıklarında 1950'lı yıllarda başlayan ve zamanla artan yönde bir trend mevcuttur. Türkiyenin iklim değişikliği çalışmalarına bilimsel bir yaklaşım getirebilmek için uyguladığımız yöntemler ve sonuçlarının ilerideki çalışmalarla birlikte bir bütün halinde düşünüldüğünde belirgin faydalar sağlayacağı görülmektedir. Daha sonraki bu tür çalışmalara yardıma olacağım gözönüne alarak verilerin zamansal değişimi ile ilgili test sonuçları da grafik ve tablolar halinde ekte sunulmuştur.

Özet (Çeviri)

Current warming trend in global temperature, tendency towards contamination of the environment especially with the man-made gases, and numerical experiment climatic models suggest that the earth climate will most probably continue to warm up in the future, (NAS, 1983). Global warming gives rise to significant changes both in local climatical features as well as hydrometeorological variables. In this study we have used the mean seasonal and annual minimum and maximum temperatures, totals of seasonal and annual precipitation. Eighteen meteorology stations are chosen for a study of climatic trend in Türkiye. These stations have a uniform geograhpical distribution, long-period records, and sosyo-economic importance. These data are analyzed in order to identify meaningful long-term trends by making use of the least square method, Spearman rank correlation and Mann-Kendall trend tests. Before using homogeneity and trend tests it is useful to analyze climatologic data for its average, median, mode, standard deviation, minimum, maximum, standardized skewness and standard kurtosis values. In this way it is possible to learn much about the data distribution and range of variability. For data homogeneity two tests are used namely, run (Swed-Eisenhart) test and serial (Wold-Wolfowitz) tests. The former is a nonparametric test used in determining the randomness or homogeneity of data set, (Swed and Eisenhart, 1943). It is used here to test the homogeneity of mean seasonal and annual minimum and maximum temperatures, seasonal and annual totals of precipitation. The number of runs of data points occurring higher than and less than the median are compared against confidence limits. If the number of runs falls between the limits, the data set considered is randomor homogeneous. However, when the number of runs falls below the lower limit, the data set is considered to have a changing mean with time. On the other hand, when the number of runs falls above the upper limit, the data set contains a high frequency oscillation. In the application of the Swed-Eisenhart test it is supposed that N consecuative values of a data set are considered for homogeneity. The maximum number of runs that could exit is A" (alternating values greater and less than the median throughout) and the minimum number two (all values greater or less than the median occurring in the first or last half of the series). If iV runs occur, an oscillation obviously exist. If the runs occur such that the greater-than and less- than values (Above and Below, respectively) are paired, a situation like BBBBAAABAAABBBAAA,...,ABBBBB exits. An oscillation is evident, but the Swed-Eisenhart test would indicate that the data points are random. One must be certain that data-set appearance and Swed-Eisenhart test results do not conflict. So that an accurate application of the test can be assured, (Kevin and Griffiths, 1985). On the other hand, the Wald-Wolfowitz test can be explained for hohogeneity as follows. If Xi,X2,...,Xn are the values of the series being considered on which a change of origin has been carried out in such a way that E^ = ° (!) and if also we put Xn+i - X\ (2) the test statistic is the quantity R = Y,Xi+Xi+1 (3) i=l whose distribution under the null hypothesis, is approximately normal for large values of n. Then mean and the variance of the distribution are fairly complex. However, if it is limited to terms of order 1/ra, this becomes xviE(R) = -S2/(n - 1) and var R = S22/(n - 1) (4) where we put S2 = JTX? (5) in addition, it is clear that if the statistic R is replaced by the function r = R/S2 (6) the mean and the variance of the asymptotic distribution becomes E(r) = -i/(n-l) and var r = l/(n - 1). (7) It must be stressed that in the serial correlation test considered, the alternative hypothesis includes only the contingency of a positive serial correlation (persistence). In order to ensure maximum power to the test, therefore, its one-sided is used, and the null hypothesis is rejected for large values of r. It follows that the test is reduced to calculation of the quantity u(r) = [(n - l)r + l]/Vn~^T (8) and of the probability a determined from a standard normal distribution table, such that a = P{u > u(r)) (9) if a0 is the significance level of the test the null hypothesis is accepted or rejected at the level a0 depending on whether a > a0 or a < ao, (Sneyers, 1990). In our data we used three tests to demonstrate the possible existence of a trend (null hypothesis). One of them is least square method which depends xviion data distribution. The other two non-parametric tests are the Spearman coefficient rs and the Mann-Kendall rank correlation statistic t. Least squares method; Let us suppose that we are fitting N data points (Xi, Yi), i = 1, 2,,3..., İV this dependence may be characterized by a relation of the form Y = f(Xu X2,..., -Yra) (10) In this case, the function / reduces to a linear function of the time; this becomes Y = AX + B (11) Here if X is zero than B is the first value of Y. However, A is the trend constant. Value and sign of A show us how values are changing with time, (Şen, 1993). The second one is Spearman coefficient rs: Calculation of this the time is i = I, 2, 3,..., N and its rank is denoted by RXi. The original observations Yi, i = 1, 2,..., N and its rank is replaced by RYi which are given to them when they are arranged in increasing order of magnitude. The correlation coefficient ra, between the RXi and RYi series, can be calculated by means of the formula r* = l-^hr)^RXi-RYi)2 (12) using the null hypothesis, the distribution of this quantify is asymptotically normal with E(rs) = 0 and var rs = (13) n - 1 It is clear that in the absence of any assumptions regarding the existence of a trend in a given direction, the test is correct only if its two-sided form is adopted, that is to say if the null hypothesis is rejected for large values of \rs\. In these conditions, after having calculated ra, it is useful to determine the probility a, by means of a standart normal distribution table, such that xvma = P(\u\>\u(rs)\) (14) where u(rs) = rs\/n - 1 (15) and the null hypothesis is accepted or rejected at the level a0, depending on weather a > aQ or a < uq. In the case of significant values of \rs\, an increasing or decreasing trend is observed depending on whether rs > 0 or rs < 0. The third one is Mann-Kendall trend test. In this test, for each element Xi or, what amounts to the same thing, for each element yi, the number n, of elements yj preceding it i > j is calculated such that yi > yj. the test statistic t is then given by equation n t = ^2m (16) i=l and its distribution function, under the null hypothesis, is asymptotically normal, with mean and variance cu\ n{n-l),, n(n - l)(2n + 5). E(t)= 4 and var * = -* ^ '- (17) It is clear that, as for the test based on the Spearman coefficient rs, in the absence of any assumption regarding the existance of a trend in a given direction, the test is correct only in its two-sided form. The null hypothesis must, therefore, be rejected for high values of u(t) with: u(i) = [t - E(t)}/Vvar t (18) In particular, if the probability a is determined using a standard normal distribution table such that a = P(u> u(t)) (19) the null hypothesis is accepted of rejected at the level a0, depending on wheather we have a > ao or a < uq. When the values of u{t) are significant, an increasing or decreasing trend can be observed depending on whether u{t) > 0 or u(t) < 0 (20) xixAfter the implimentation of homogeneity and trend tests it is concluded that most of the data are homogeneous. The data of some stations are inhomogeneities for both tests. These stations are; for minimum temperature data, Konya, Urfa, Diyarbakır and Antalya in spring. Sivas, Çanakkale, Van, Diyarbakır, Antalya and Adana in summer. Ankara, Elazığ and Antalya in autumn. Sivas, Antalya and Adana in annual average. For maximum temperature, Diyarbakır in annual average. Göztepe, Urfa and Antalya in winter. Ankara and Konya in spring. Konya in summer. For precipitation data, Zonguldak in spring. Van in autumn. There is a big discontinuty at two instances Antalya data in 1937 and 1974. Generally there is no trend in precipitition and maximum temperature in Türkiye, but there is increasing trend in minimum temperatures. There is in increasing and decreasing way in some stations for both Spearman and Mann-Kendall test. These stations are; for minimum temperature, Kars and Adana's temperatures are increasing in winter. Zonguldak, Göztepe, Sivas, Kars. Çanakkale, Kütahya, Van, Elazığ, İzmir, Konya, Ş. Urfa, Diyarbakır and Adana's temperatures are increasing in spring. Zonguldak, Edirne, Göztepe, Sivas, Çanakkale, İzmir, Konya, Ş. Urfa, Diyarbakır and Antalya's temperatures are increasing in summer. Trabzon, Sivas, Ş. Urfa and Adana's temperatures are increasing in autumn. Zonguldak, Göztepe, Sivas, Kars, Çanakkale, Kütahya, Van, Elazığ, İzmir, Konya, Ş. Urfa, Diyarbakır and Adana's temperatures are increasing in annual data. For maximum temperature, Kütahya temperature is increasing in spring. Van and Ş. Urfa in summer, Samsun, Van and Ş. Urfa's temperatures are increasing in autumn. For precipitation Samsun and Kars in winter, Kars and Ankara's precipitation are decreasing. Ankara in spring, Antalya in summer, Ankara in annual precipitation is increasing. xx

Benzer Tezler

  1. Tez No

    688572

    The impact of climate change on future extreme precipitation in Turkey

    Türkiye'de iklim değişikliğinin gelecek ekstrem yağışlara etkisi

    DERYA ALBAYRAK

    Yüksek Lisans

    İngilizce

    İngilizce

    2021

    Bilim ve Teknolojiİstanbul Teknik Üniversitesi

    İklim ve Deniz Bilimleri Ana Bilim Dalı

    PROF. DR. ÖMER LÜTFİ ŞEN

  2. Tez No

    142736

    Climate variability Jordan

    Ürdün'de iklim değişikliği

    MUWAFFAQ FREİMAN

    Yüksek Lisans

    İngilizce

    İngilizce

    2003

    Meteorolojiİstanbul Teknik Üniversitesi

    Meteoroloji Mühendisliği Ana Bilim Dalı

    PROF. DR. MİKDAT KADIOĞLU

  3. Tez No

    493269

    Romatizmal hastalıklarda, tekrarlanan kaplıca tedavisi kürlerinin etkinliğinin retrospektif değerlendirilmesi

    The retrospective evaluation of repeated spa cure therapies in rheumatologic diseases retrospective evaluation of the effectiveness of the repeate treatments cylinder in the rheumatic diseases

    ERGİCAN KERİM ESİN

    Doktora

    Türkçe

    Türkçe

    2016

    Tıbbi Ekoloji ve Klimatolojiİstanbul Üniversitesi

    Tıbbi Ekoloji ve Hidroklimatoloji Ana Bilim Dalı

    PROF. DR. HATİCE GÜRDAL

    PROF. DR. MÜFİT ZEKİ KARAGÜLLE

  4. Tez No

    554046

    Titriş Höyük (Bahçeli / Karaköprü / Şanlıurfa) ve yakın çevresinde arkeocoğrafya araştırmaları

    Archaeogeography researches on Titriş Höyük (Bahçeli / Karaköprü / Şanliurfa) and immediate surroundings

    MEHMET BUĞU

    Yüksek Lisans

    Türkçe

    Türkçe

    2019

    Arkeolojiİstanbul Üniversitesi

    Coğrafya Ana Bilim Dalı

    PROF. DR. HÜSEYİN TUROĞLU

  5. Tez No

    557662

    Low flow duration frequency curves and trend analysis

    Düşük akım süre frekans eğrileri ve trend analizi

    SELEN ORTA

    Yüksek Lisans

    İngilizce

    İngilizce

    2019

    Çevre Mühendisliğiİstanbul Teknik Üniversitesi

    İnşaat Mühendisliği Ana Bilim Dalı

    PROF. DR. HAFZÜLLAH AKSOY

Geri Dön