as a PDF
Transkript
as a PDF
World Applied Sciences Journal 12 (10): 1662-1675, 2011 ISSN 1818-4952 © IDOSI Publications, 2011 Analysis of Spatially Distributed Annual, Seasonal and Monthly Temperatures in Istanbul from 1975 to 2006 Ahmet Karaburun, Ali Demirci and Fatih Kara Faculty of Arts and Science, Geography Department, Fatih University, Turkey Abstract: The present study is about the analysis of mean, mean maximum and mean minimum temperatures carried out on annual, seasonal and monthly time-scales examining the data from 8 meteorological stations in Istanbul for the period 1975-2006. Various spatial and statistical tools were used to display and analyze trends in temperature data. ArcGIS was used to produce the spatially distributed temperature data by using Thiessen method. The non-parametric Mann-Kendall test was used to determine whether there is a positive or negative trend in data with their statistical significance. Sen’s method was also used to determine the magnitude of the trends. The results reveal positive trends in annual mean and mean maximum temperatures with 95% and 99% significance. Annual mean temperatures increased 0.83°C while annual mean maximum temperatures increased 1.6°C over the 32-years period in Istanbul. On a seasonal basis, a statistically significant positive trend was observed in summer. On a monthly basis, July and August has experienced significant warming trend in mean, maximum and minimum temperatures. The analysis of the whole record reveals a tendency towards warmer years, with significantly warmer summer, spring and autumn periods and slightly colder winters. These warming patterns may have important impacts on energy consumption, water supply, human health and natural environment in Istanbul. Key words: Thiessen polygon Temperature Mann-Kendall INTRODUCTION The world has been witnessing the effects of global warming in many different areas during the last century. Rising sea levels, glacier and snow cover retreat and changing patterns of agriculture are among the many consequences which are attributed to global warming [1-4]. The increase in global surface temperature and its direct and indirect effects have already caused many widespread social, economic and environmental problems throughout the world. Changing patterns and timing of extreme weather events such as tropical cyclones, expansion of contagious diseases and increasing duration of heat waves have been cited among negative effects of global warming on human life in many studies [5-8]. Studies on monitoring global and regional temperature change have intensified over the last few decades because global and regional effects of global warming became apparent. One of the most important associations providing data on global temperature changes is the Intergovernmental Panel on Climate Change (IPCC) which is a scientific intergovernmental body established in 1988. According to IPCC’s recent report, the global average surface air temperature has increased by 0.13 °C (± 0.03°C) per decade over the last 50 years [9]. As estimated in many studies, global temperature increased 0.8°C in the past century and the warming was not continuous but occurred principally during 1920-45 and after 1975 [10,11,12]. Cooling was significant during the intervening period (1945-75) especially for North America, the Arctic and Africa [10]. Analysis of worldwide air temperature changes have also shown that temperature has increased in both hemispheres over the last century; however, warming was more dominant in the Northern Hemisphere since the 1950s [12]. Many regional studies have also found a positive trend in temperature although the changes slightly vary from one region to another [13-21]. As Zhang et al. [22] indicated, the annual mean temperature has increased between 0.5 and 1.58°C in the southern regions of Canada in the 20th century. A strong warming trend of +0.08°C ± 0.03°C per decade was detected over Europe within the 20th century with the highest increase as 0.43°C over the last 30 years [14]. Corresponding Author: Ahmet Karaburun, Faculty of Arts and Science, Geography Department, Fatih University, Turkey. Tel: +90-212-8663300(2038), Fax: +90-212-8663402, E-mail: akaraburun@fatih.edu.tr. 1662 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Analysis of temperature data throughout the world has shown that change was not only observed in mean annual temperature but also in seasonal, monthly and minimum-maximum temperatures. As Pielke et al. [23] stated, minimum temperatures have risen about 50% faster than maximum temperatures globally since 1950. Winter minimum and maximum temperatures significantly increased in USA, Europe and many parts of Asia while small increase was observed on minimum and maximum temperatures during the summer [24]. The highest increases in temperature were found in winter in the interior region of Alaska as 2.2°C within the last 50 years [22]. Mediterranean region is among the most affected areas in the world from global warming. Studies on global warming in the Mediterranean basin indicate that the rise in temperature which intensified over the last few decades increased the potential evapotranspiration, water deficit and forest fire risk [25]. Brunet et al. [17] studied temperature variability in Spain during 1850-2005 and found a significant warming of 0.10°C per decade for the annual mean temperature. According to the same study, average autumn and winter temperatures have increased slightly more than spring and summer temperatures. Founda et al. [15] studied mean, maximum and minimum temperature in Athens from 1897 to 2001 and concluded that an increase of 1.23 and 0.34°C has been observed in the mean summer and mean winter temperature respectively. A significant increase has also been observed for both warm and cold seasons in the same study between 1992 and 2001. A number of studies used different methods and predicted the changes in average, maximum and minimum temperature in Turkey with changing patterns from region to region [26-32]. Turke et al. [26] had found cooling in many stations of Turkey during the second half of the 20th century. However, as Tayanc et al. [31] indicated, almost all parts of Turkey experienced a warming trend over the last half century except for the northern parts. A more widespread significant warming trend was reported in the minimum temperatures in the southern, western and continental parts of Turkey in the same study. Black Sea region, however, has experienced a cooling period from 1975 to 1993, but a warmer period from 1994 to 2007 [30]. As seen in some recent studies, predicting the change in annual, seasonal mean, maximum and minimum temperatures for the whole Turkey is problematic since different trends are observed in different climatic regions of the country [26,30,31,33]. For this reason, mean, maximum and minimum temperature changes have to be analyzed regionally in Turkey by using different methods in order to understand the trends in temperature and make correct predictions for future temperature changes. Understanding of climate change requires a careful examination and interpretation of climatic data to see if there is a change in the average of the data in a given time period. Several parametric and non-parametric statistical data processing methods are used to analyze trends in climate change studies. Sequential Mann–Kendall trend test, Sen’s slope estimator and Spearman’s rank–order correlation tests are used to analyze the direction and magnitude of possible trends in temporal observed data [55, 56, 57, 58]. Many researchers used Mann-Kendall tests to identify trends in hydrologic variables such as river flow, sediments and meteorological variables such as temperature and precipitation in many studies [59, 60, 61, 62, 29, 63]. The purpose of this study is to analyze trends in annual, seasonal and monthly mean, minimum and maximum temperatures in Istanbul for 32 years time period (1975-2006) by using Mann-Kendall non-parametric test and Sen’s method applied on spatially distributed monthly temperature data produced by Thiessen polygon method. Study Area: Study area is the financial capital of Turkey, Istanbul, with its population over 12 million [64]. The city is located in northwest Turkey along both sides of the Bosphorus, where the western side of the city is in Europe and its eastern side is in Asia (Fig. 1). It is located between 40° 48'– 41° 36' N latitudes and 27° 58'–29° 56' E longitudes. Istanbul covers a surface area of 5,390 km2 with 32 sub-provinces administered by the Istanbul Metropolitan Municipality. Istanbul is surrounded by Black Sea in the north and Sea of Marmara in the south. These two water bodies along with the Mediterranean Sea determine the climatic characteristics of the region. The climate of the city shows a transition between the Mediterranean and temperate climates characterized by cool and wet winters and warm and humid summers [65]. The average annual temperature is 13.7°C with daily mean temperature 6.3°C in winter and 22.4°C in summer [66]. According to data taken Florya Meteorological station which is located in the south, the hottest months are July (23.2°C) and August (23.7°C) while the coldest months are January (5.3°C) and February (5.5°C) [66]. Total annual precipitation is 734 mm in the city [67]. Like temperature, the amount of precipitation varies from south to north. 1663 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Fig. 1: Study area map The total annual precipitation is over 1.000 mm in the north while it decreases towards Sea of Marmara in the south below 600mm [67]. The topography in European and Asian sides of the city is characterized by low plateaus. Average elevation of the area is between 150 and 200 meters, though some mountains are higher than 500 meters. The elevation decreases from north to south in general over Istanbul. Rapid increase in population and urbanization has resulted significant changes in land cover of the city especially over the last fifty years. Being located mostly in the northern parts, forest areas cover 37.3% of the total surface are of the city while urban areas account for 14.3% [68]. Methodology: Estimating the spatial distribution of climatic data has become an important part of studies helping to understand climate change and its effects throughout the world [34]. While many of the studies analyze climatic data in local meteorological stations, spatial distribution of climatic data is produced by estimating data at non monitored locations based on registered values at neighboring sites [34]. Different methods were proposed in several studies to estimate spatial distribution of climatic data mainly precipitation and temperatures [36-43, 34]. The methods which are commonly used to present spatial distribution of climatic data can be classified into three main groups; graphical, topographical and numerical [42]. Graphical methods include precipitation-elevation analysis, isohyets mapping, [44, 37] and Thiessen’s [36]. Topographical methods include the correlation of point climate data obtained from topographic and synoptic parameters such as slope, exposure, elevation, the location of barriers, wind speed and direction [45, 46, 38, 47]. Numerical methods are the most attempted methods in order to estimate spatial distribution of climate data over the last years. Optimal interpolation [48], kriging and its variants [41] and smoothing splines [39] are examples in numerical methods. In Thiessen method, each meteorological station is weighted in direct proportion based on its area in the total area of region without consideration of topography or other local physical characteristics. In this method, it is accepted that the area defined by each meteorological station is closer to that station than to any other station over the study area. The Polygons that represent the influence area for each station are created by connecting the intersections of perpendicular bisectors of lines that are drawn between stations. The spatial distribution of climatic data is calculated by multiplying the climate data of each station with their areal ratio over total area. Many researchers used Thiessen method to estimate the spatial distribution of precipitation and temperature in several studies [49-53]. This method can be used for the areas which have insignificant topographic differences and inadequate meteorological station density. 1664 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Table 1: Characteristics of meteorological stations Fig. 2: Thiessen Polygon over study area Several spatial and statistical methods were used in the study to produce spatially distributed temperature data and to detect the trend on annual, seasonal and monthly mean, maximum and minimum temperatures in Istanbul for a 32-years period from 1975 to 2006. Monthly temperature data were obtained from 8 meteorological stations one of which is located in the west, outside of the provincial border of the city (Fig 2). Seven meteorological stations located in the city were those from which consistent temperature data were available for a 32-years period [54]. The only station which is located outside of the city was used to create Thiessen polygons in the western part of the city where no other stations were available for the study. Table 1 displays the location and elevation of these meteorological stations. As seen from the table, elevation is below 200 meters in every station. This was helpful to use Thiessen method to produce spatially distributed temperature data over the study area. After Thiessen method, spatially distributed data were analyzed by Mann-Kendall test and Sen’s Method for trend analysis. Producing Spatially Distributed Temperature Data: In Thiessen method, the area is divided into a set of polygons with the meteorological station in the middle of each polygon without taking into account topography and physical characteristics of study area. All points that fall inside of a Thiessen polygon are closer to center of that polygon than to center of neighbor polygons in study area [69]. The centers of polygons are represented by meteorological stations. Temperature value of each Thiessen polygon is assumed to be constant and equal to value of its representative meteorological station in Station Name Latitude Longitude Elevation Bahcekoy 41 10 35 28 59 35 130 Corlu 41 09 25 27 49 05 183 Florya 40 58 55 28 47 20 37 Goztepe 40 58 42 29 03 21 32 Kartal 40 53 23 29 10 57 27 Kirecburnu 41 08 51 29 03 02 58 Kumkoy 41 15 04 29 02 19 38 Sile 41 10 11 29 36 03 83 the center of Thiessen polygon. Thus this method allows estimating temperature over a large area using limited numbers of meteorological stations. This method is probably one of the most used method to model the spatial distribution of climatic data [49, 50, 51, 52, 53]. Discrete change in data because of discontinuous surface that occur at the boundaries of polygons is the negative effect of this method [70]. The spatially distributed temperature is estimated by determining the weights of each Thiessen polygon over the study area. The area percentage of each Thiessen polygon is used to determine the weights. Firstly, triangles are created by connecting stations with lines. Secondly, perpendicular lines are created for each bisector of lines of triangles. Those perpendicular lines will intersect and form a vertex of Thiessen. Thiessen polygon is created using those vertexes. The weights for Thiessen polygons of study area computed using the following equation; Wi = Ap A (1) Wi : The weighted area for Thiessen polygon AP : The area defined by Thiessen polygon A : The total area Spatial weighted temperature of each meteorological station is calculated by multiplying its spatial weight by its observed temperature value. The average temperature of study area is calculated by adding all aerial weighted temperature of meteorological stations over study area using following equation; P= n ∑ i =1Wi Pi (2) P : The average temperature of study area Pi : The meteorological station temperature for each Thiessen polygon n : The number of Thiessen polygon over study area 1665 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Table 2: Station Weights Station Weight CORLU 0.196 FLORYA 0.326 BAHCEKOY 0.086 KUMKOY 0.011 KIRECBURNU 0.073 SILE 0.164 KARTAL 0.107 GOZTEPE 0.037 All geographical information system software offer tools to create Thiessen polygons. The locations of meteorological stations are created using geographical coordinates provided by Turkish State Meteorological Service. The Thiessen polygons (Fig.2) that represent the influence area of each meteorological station were produced using Arcgis 9.2 software. The weights of stations are given in Table 2. Since Sen’s slope estimator is not affected from gross data error or outliers, it is used to estimate the slope for Mann-Kendall test even with missing data. Possible slopes between all possible data pairs can be calculated by Sen’s slope estimator. The initial value of the MannKendall test statistic S is set to 0 and it indicates there is no trend. It increases the S value by 1 when a laterobserved value is higher than earlier-observed value and decreases by 1 when a later-observed value is lower than an earlier-observed value. The final value of S is computed by the net results of those increases and decreases. Mann-Kendall method firstly calculates S statistic which indicates the sum of the difference between the data points indicated in following equation. = S n −1 n ∑ ∑ sgn(x j − xk ) k= 1 j= k +1 (3) Where; xj is observed value at time j, xk is observed value at time k, j is time after time k and n is the length of the data set. The sign of the value is defined as follows: Determining the Trend Analysis of Spatially Distributed Temperature Data: Spatially distributed temperature data obtained from Thiessen method was grouped into annual, seasonal and monthly mean data series in the period of 1 if x j - xk > 0 1975-2006 to be used in Mann-Kendall test and Sens’ method. Before applying those test to temperature data = sgn (x j -xk ) = 0 if x j - xk 0 (4) series time period was divided into two periods. The -1 if x x 0 < j k Mann-Kendall test was employed to detect presence of positive or negative trend in time series data while Sen’s The statistical significant of S is checked using a test method was used to compute the magnitude of the statistic (or z score): change in temperature. S −1 One of the most used non-parametric tests to Var(S ) ; S > 0 identify presence of trends in time series data is The (5) Mann-Kendall test. Missing values in time series data = Z = 0 ; S 0 S +1 are allowed in Mann-Kendall trend test and the data do ; S < 0 not require conforming to any particular distribution Var(S ) [71]. The sign of difference is identified after all subsequent observed values are compared and the The variance of S can be computed as: difference between the later-observed values and earlierq observed values are compared in the Mann-Kendall test. n( n − 1)(2n + 5) t p (t p − 1)(2t p + 5) Each-later observed value is analyzed with all earlieri =1 (6) Var(S ) = observed values. If later-observed values tend to be larger 18 than earlier-observed values, an increasing trend can be Where n indicates the number of data points, q indicates identified while a decreasing trend can be recognized if the number of tied groups (a tied group is a set of sample later-observed values tend to be smaller than earlierdata having the same value) and tp indicates the number observed values. of data points in the pth group. Linear regression method is used to estimate the The Mann-Kendall method test the null hypothesis, slope of possible linear trend in time series but if there are Ho, is that data indicate no distinct trend against the gross data errors or outliers in time series, the estimated alternative hypothesis Hi is that data indicate a trend. slope can be different from the true slope of linear trend. ∑ 1666 World Appl. Sci. J., 12 (10): 1662-1675, 2011 RESULTS Two tailed test is used to decide whether or not the null hypothesis should be rejected in favor of hypothesis test. The null hypothesis, Ho is tested by the Z test statistic value. A decreasing trend is identified if Z is negative and an increasing trend is identified if the Z is positive. Ho is rejected at 0.1, 0.05, 0.01 and 0.001 significance levels if the absolute value of Z is greater than Z1- /2, where Z1- /2 is taken from the standard normal cumulative distribution tables. The magnitude of the trend over time is estimated according to Sen [72]. The median slope of all pair wise comparisons indicates the trend slope and the slopes of all data pairs are calculated by; Qi = xi - x k j-k Spatially distributed temperature data which were produced by using Thiessen method and their analysis with Mann-Kendall test and Sen’s method provided an indepth analysis of annual, seasonal and monthly mean, maximum and minimum temperature changes in Istanbul during the last 32 years (1975-2006). (7) Where Qi : Slope between data points xj : Data measurement at time j, xk : Data measurement at time k j > k. If there are n values of xj in the time series Sen’s slope estimator is the median of n(n-1)/2 pairwise slopes. The Sen's estimator is: Q = Q N +1 2 if N is odd, (8) Annual Temperatures: Table 2 presents the spatially distributed annual mean, maximum and minimum temperatures. The analysis of the whole dataset within the table indicates that annual mean temperature is 13.7°C in Istanbul. As table 3 indicates, mean annual temperatures varied between 12.6°C and 14.9°C in Istanbul from 1975 to 2006. The mean annual temperature was the lowest in 1976 while it was the highest in 2001 with 2.3°C difference between the two. If the data in table 3 are examined in two different time periods (1975-1990 and 1991-2006), annual mean temperatures seem to have differed slightly. Annual mean temperature is 13.5°C in the first period while it is 13.9°C in the second period. This result reveals that annual mean temperatures increased especially after 1990 (Figure 3). Before comparing the mean, minimum and maximum temperature data of two time periods, the Student’s t test was applied to assess temperature differences between two time periods whether they are statistically different from each other. A statistically significant difference has 1 (9) Q N + Q N + 2 if N is even. 2 2 2 Table 3: Spatially distributed annual mean, maximum and minimum temperatures in Istanbul from 1975 to 2006 Q = Temperatures (°C) Temperatures (°C) ----------------------------------------------------------------- ------------------------------------------------------------------- Year Mean Max. Min. Year Mean Max. Min. 1975 13,8 24,4 5,1 1991 13,2 23,8 5,0 1976 12,6 23,4 3,6 1992 13,1 25,6 4,1 1977 13,7 25,7 4,7 1993 13,1 26,0 4,4 1978 13,5 24,5 4,9 1994 14,7 26,8 5,5 1979 14,1 25,5 4,1 1995 14,0 25,0 4,9 1980 13,2 24,6 4,2 1996 13,3 24,3 5,2 1981 13,6 24,2 4,5 1997 13,1 25,1 3,9 1982 13,2 24,8 4,4 1998 14,1 25,6 5,1 1983 13,4 24,8 4,2 1999 14,8 25,3 5,6 1984 13,6 25,1 4,8 2000 14,2 26,3 4,6 1985 13,1 25,1 3,4 2001 14,9 26,2 5,3 1986 13,5 23,9 4,4 2002 14,5 26,3 5,6 1987 12,9 25,1 2,7 2003 13,6 25,8 4,5 1988 13,4 24,6 4,2 2004 14,0 26,0 3,8 1989 13,7 25,6 4,9 2005 13,9 24,6 4,1 1990 14,1 25,6 4,6 2006 13,9 25,0 4,6 1667 World Appl. Sci. J., 12 (10): 1662-1675, 2011 16,00 Tmean 15,50 15,00 14,50 14,00 13,50 13,00 12,50 12,00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Years Tmax 27,00 26,50 26,00 25,50 25,00 24,50 24,00 23,50 23,00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Years Fig. 3: Annual Tmean, Tmax, Tmin and their linear trends for Istanbul between 1975 and 2006. Table 4: Annual and seasonal temperature trends of Istanbul over the 32-years period by using Mann-Kendall test and Sen’s method (spring: March-May; summer: June-August; autumn: September-November; winter: December-February) Tmean (°C) Tmean, max (°C) Tmean, min (°C) Spring 0.67 1.15 -0.32 1.92** Summer 2.24*** 1.73* Autumn 0.83+ 1.38 0.51 Winter -0.16 0.16 -0.48 Year 0.83* 1.6** 0.54 + Significance level <= 90%. *Significance level <= 95%. **Significance level <= 99%. ***Significance level <= 99.9% been found at 0.05 level of significance between data series displayed in two time periods as a result of the Student’s t test. According to the same data presented in Table 3, mean annual maximum and minimum temperatures are 25.1°C and 4.5°C respectively in Istanbul. Mean annual maximum and minimum temperatures display more variability between years in comparison with annual mean temperatures. The lowest mean annual maximum temperature was observed in 1976 as 23.4°C while the highest value was seen in 1994 as 26.8°C with 3.4°C difference between the two. Comparison of the mean annual maximum temperatures within two time periods (1975-1990 and 1991-2006) reveals that there is 0.7°C difference between the two. The mean annual maximum temperature was found to be 24.8°C in the first period from 1975 to 1990 while it increased to 25.5°C in the second time period from 1991 to 2006. The mean annual minimum temperatures also display similar pattern between years although the increase in values slightly lower than the mean annual maximum temperatures. The lowest mean annual minimum temperature was observed in 1987 as 2.7°C while the highest value was seen in 1999 and 2002 as 5.6°C. Analysis of the mean annual minimum temperature within the two time periods reveals that the second period (1991-2006) experienced a more warming than the first period (1975-1990). The mean annual minimum temperature was 4.3°C in the first period (19751990) while the same figure was 4.8°C in the second period (1991-2006). The application of the Mann-Kendall test and Sen’s method demonstrated a positive trend in annual mean, maximum and minimum temperatures over the 32-years period in Istanbul. The trends in annual mean and maximum temperatures are significant at the 95% and 99% respectively while annual mean minimum temperature is not statistically significant (Table 4). As seen from the table 3, the annual mean warming trend in Istanbul was 0.83°C during a 32-years period (1975-2006). This result indicates a warmer trend in the study area than the average annual increase in global temperatures which was 0.13 °C (± 0.03°C) per decade over the last 50 years [9]. 1668 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Table 4: Spatially distributed seasonal mean, maximum and minimum temperatures in Istanbul over the 32-years period (spring: March-May; summer: JuneAugust; autumn: September-November; winter: December-February) Spring Summer Autumn Winter Tmean (°C) Tmean, max (°C) Tmean, min (°C) 11,5 22,3 15,1 5,9 25,0 32,7 26,4 16,5 2,0 13,4 6,0 -3,3 Table 5: Spatially distributed monthly mean, maximum and minimum temperatures in Istanbul over the 32-years perio Months Tmean (°C) Tmean, max (°C) January 5,3 15,6 February 5,1 17,1 March 7,1 21,0 April 11,5 25,5 May 16,0 28,5 June 20,7 32,3 July 23,1 33,0 August 23,0 32,8 September 19,5 30,6 October 15,2 27,0 November 10,5 21,5 December 7,2 16,9 Annual 13,7 25,1 Although there is a positive trend in annual mean maximum and minimum temperatures, trend is more important for maximum temperatures. Annual mean maximum temperature increased 1.6°C, while annual mean minimum temperature experienced 0.54°C increase over the period between 1975 and 2006 (Table 4). As this result reveals, warming trend in annual mean maximum temperature has played an important role in the increase in annual mean temperature. Fig. 3 shows annual Tmean, Tmax, Tmin trends for Istanbul on over the period 1975-2006 using Sen’s model. As clearly seen from the figure, the warming trend is more obvious in annual Tmax altough annual Tmean and Tmin have also experienced positive trends. Seasonal Temperatures: Analysis of seasonal temperature data provided a better understanding of warming trend in Istanbul. Table 5 represents spatially distributed seasonal mean, maximum and minimum temperatures in Istanbul over the 32-years period. As the table reveals, mean temperatures were 11.5°C in spring, 22.3°C in summer, 15.1°C in autumn and 5.9°C in winter. Mean maximum temperatures varied from 16.5°C in winter to 32.7°C in summer. Mean maximum temperatures in spring and autumn were 25°C and 26.4°C respectively. Mean minimum temperatures dropped below zero as 3.3°C in winter while it was 13.4°C in summer. Spring and autumn have experienced slightly low mean minimum temperatures as 2°C and 6°C respectively in the same period. Tmean, min (°C) -3,7 -4,1 -2,0 2,0 6,1 11,1 14,5 14,6 10,7 6,2 1,0 -2,1 4,5 The analysis of the evaluation of T mean by seasons with Mann-Kendall test and Sen’s method indicates a positive trend in all seasons except winter in Istanbul over the period between 1975 and 2006. The trends in T mean for spring and winter are not statistically significant. However, the trends in summer and autumn are significant at the 99.9% and 90% respectively (Table 4). As table 4 reveals, summer is the season that experienced the highest warming trend in T mean. Over the 32-years period from 1975 to 2006, T mean has increased 2.24°C in summer in Istanbul. Spring and autumn have also experienced a warming over the same time period as 0.67°C, 0.83°C respectively. Although the trend is not statistically significant, winter has experienced 0.16°C decrease in Tmean over the same period in Istanbul. Fig. 4 shows seasonal Tmean and their linear trends for Istanbul from 1975 to 2006 using Sen’s model. Highest warming trend in summer and slight cooling trend in winter is obvious in Tmean from the figure. Mann-Kendall test has revealed a positive warming trend in Tmax in all seasons in Istanbul for the 32-years period (Table 4). Warming trends in T max were found statistically insignificant in all seasons except summer. The trend in Tmax in summer is significant at 95%. As the results in table 4 reveals, summer was the season experienced the highest warming in Tmax as 1.73°C over the 32-years period in Istanbul. Spring and autumn have also experienced warming trends in T max over the same priod as 1.15°C and 1.38°C respectively. Winter did not experience a significant increase in Tmax so it only increased 0.16°C in 1669 World Appl. Sci. J., 12 (10): 1662-1675, 2011 15,00 Spring 25,00 14,00 24,00 13,00 23,00 12,00 22,00 11,00 21,00 10,00 20,00 9,00 19,00 8,00 1970 1975 1980 1985 1990 1995 2000 18,00 1970 1975 1980 2005 2010 Summer 1985 1990 Years 1995 2000 2005 2010 Years Fig. 4: Seasonal Tmean and their linear trends for Istanbul between 1975 and 2006. 32 years. A different trend was observed in Tmin for seasons. As table 4 reveals, spring and winter have experienced negative trends in Tmin over the 32-years period although the trends were not statistically significant. Tmin has decreased 0.32°C and 0.48°C over the same period in spring and winter respectively in the city. The analysis of the evolution of Tmin by seasons indicates 1.92°C increase in summer over the periods of 32-years with 99% significance level. Although it is not statistically significant, autumn has also experienced 0.51°C increase in Tmin over the same period in Istanbul. Monthly Temperatures: Table 5 displays the spatially distributed monthly mean, maximum and minimum temperatures for Istanbul over the period from 1975 to 2006. As seen from the table, the July and August are the warmest months in Istanbul with temperatures 23.1°C and 23°C respectively. According to the mean temperatures, January and February are the coldest months in Istanbul with temperatures 5.3°C and 5.1°C respectively. The highest monthly mean maximum temperatures were observed in July with 33°C while the same figure was slightly lower in August as 32.8°C. The lowest monthly mean maximum temperatures were experienced in January with 15.6°C. Mean minimum temperatures show a similar pattern. July and August experienced the highest mean minimum temperatures as 14.5°C and 14.6°C respectively while the same figure was the lowest in February as 4.1°C. Table 6 presents the results of applying the MannKendall test and Sen’s method on monthly mean, maximum and minimum temperatures of Istanbul over the 32-years period. It was observed that mean temperatures have experienced a positive trend in all months except November and December. The most important T mean increases were detected in July and August with 99.9% significance. August experienced the highest increase in Tmean as 2.94°C while the increase was 2.37°C in July from 1975 to 2006. A significant warming trend in Tmean was also observed in May, June and September (Table 6). T mean has increased 1.25°C in May, 1.18°C in June and 1.12°C in September from 1975 to 2006. Although summer months experienced a significant warming trend, November and December had a cooling trend in Tmean over the 32-years period in Istanbul. Although the trend is not statistically significant, Tmean decreased 0.26°C in November and 0.38°C in December in Istanbul from 1975 to 2006. A quite small increase was detected in Tmean in January and March although the trend was statistically insignificant. T mean has only increased 0.03°C in January and 0.10°C in March over the 32-years period in Istanbul. Results of Mann-Kendall test and Sen’s method revealed that the trend in monthly T max and T min is different from that of Tmean. As seen from the table 6, all months except March demonstrated a positive trend in Tmax. The most important warming trend in T max was observed in August with 99.9% significance. Over the 32-years period from 1975 to 2006, Tmax increased 4.13°C in August in 1670 World Appl. Sci. J., 12 (10): 1662-1675, 2011 Table 6: Monthly temperature trends of Istanbul over the 32-years period (1975-2006) by using Mann-Kendall test and Sen’s method Months January February March April May 1.25+ June 1.18* July 2.37*** August September October November December + Tmean (°C) 0.03 0.38 0,10 0.45 2.18 0.64 0.67 2.94*** 1.12+ 1.38 -0.26 -0,38 Tmean, max (°C) 0.90 1.38 -0.29 1.50 -0,10 1.76+ 1.76* 4.13*** 2.08* 0.86 1.18 0.58 Tmean, min (°C) 0.35 0.13 0.26 -1.50 2.05** 1.41 0.90 -0.99 -1.76 Significance level <= 90%. *Significance level <= 95%. **Significance level <= 99%. ***Significance level <= 99.9% 26,00 Tmean 39,00 25,00 37,00 24,00 23,00 35,00 22,00 33,00 21,00 31,00 20,00 29,00 19,00 1970 Tmax 1980 1990 Years 2000 27,00 1970 1975 1980 1985 1990 1995 2000 2005 2010 2010 Years 18,00 Tmin 17,00 16,00 15,00 14,00 13,00 12,00 11,00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Years Fig. 5: Tmean, Tmax, Tmin and their linear trends in August for Istanbul between 1975 and 2006 Istanbul. Another striking result revealed from Sen’s method is about the increase in Tmax in September. Tmax increased 2.08°C in September with 95% significance while June and July have experienced a lower warming trend as 0.64 and 0.67 respectively over the 32-years period. Although they had a negative trend in Tmean, November and December experienced a warming trend in Tmax as 1.18°C and 0.58°C respectively over the same period. Although statistically insignificant, increase in Tmax in January and February is much pronounced than Tmean. These two months have experienced 0.90°C and 1.38°C increase in their Tmax respectively over the same period (Table 6). March was the only month that experienced a cooling trend in Tmax. Tmax decreased 0.29°C over the same period in March. The analysis of the evalution of Tmin by months with Mann-Kendall test and Sen’s method indicates a negative trend in April, May, November and December while other months have experienced a positive trend over the period between 1975 and 2006. The trends in T min are not statistically significant in the months except June, July and August. The trends in these months are significant at 90%, 95% and 99% respectively. August experienced the highest increase in Tmin as 2.05°C over the 32-years period. As seen from the fig. 5, T mean, T max and T min have increased in August with over 99% significance. June and July have also demonstrated a 1.76°C increase in their T min over 32years period. In contrast to its T mean and T max, T min has displayed a cooling trend in April and May. Although T min decreased 1.50°C in April, the cooling was only 0.10°C in 1671 World Appl. Sci. J., 12 (10): 1662-1675, 2011 May over the same period. As the results reveal, November and December had experienced a significant decrease in Tmin over 32-years period as 0.99°C and 1.76°C respectively. CONCLUSION The present study analyses the evolution of annual, seasonal and monthly mean, minimum and maximum temperatures in Istanbul for a 32-years period between 1975 and 2006 by using Mann-Kendall test and Sen’s method applied on spatially distributed monthly temperature data produced by Thiessen polygon method. The results presented in this paper reveal positive trends in annual Tmean, Tmax and Tmin. Annual Tmean has increased 0.83°C while Tmax and Tmin had experienced an increase of 1.6°C and 0.54°C respectively in Istanbul over the 32-years period. As these results indicate, Istanbul has experienced a higher warming trend than many parts of the world especially the European continent in annual Tmean over the last three decades. The increase in Tmean were reported as 0.8 °C in the world in the last century [10] and as 0.43°C in Europe over the last 30 years [14]. The annual Tmean has increased nearly double in Istanbul in comparison with European continent over nearly the same time period. The analysis of temperature evolution carried out on a seasonal basis reveals a stronger statistically significant positive trend in summer in Tmean, Tmax and Tmin. Tmean has increased 2.24°C in summer while Tmax and Tmin have experienced the increase as 1.73°C and 1.92°C respectively in the same 32-years period. Spring and autumn have also experienced a significant warming trend in Tmean, Tmax and Tmin except spring having a negative trend in its Tmin. In contrast to warming trend found in other seasons, winter has demonstrated a negative trend over the same period. Although it is not statistically significant, Tmean had decreased 0.16°C in winter. The change on seasonal temperatures over the last few decades has been observed differently in many parts of the world. Although a warming trend has been reported on summer temperatures of different regions in other studies [24], 2.24°C increase on summer annual mean temperatures is striking in Istanbul. The negative trends observed in this study on winter temperatures of Istanbul are also quite different from what was measured in other areas around the world [15, 22]. This indicates that global and regional studies on global warming should not be used alone and should be evaluated together with local studies to predict how climate will change in local areas. The analysis of monthly temperature data in Istanbul has provided a better understanding of the change in Tmean, Tmax and Tmin over the 32-years period. The most important warming trend was observed in August in Tmean, Tmax and Tmin with statistical significance over 99%. Tmean has increased 2.94°C in August while Tmax and Tmin had experienced the increase as 4.13°C and 2.05°C respectively. July has also experienced a significant warming trend in Tmean with 99.9% significance over the same period. Although it is not statistically significant, November and December have experienced a decrease in Tmean as 0.26°C and 0.38°C over the period from 1975 to 2006. The study clearly shows very strong temperature increase in summer and a little decrease in winter in Istanbul during the last 32 years (1975-2006). The study revealed that the warming trend especially observed in the northern hemisphere over the last half century was also experienced in Istanbul. The results of the study comply with other researches which indicated a warming trend all over Turkey especially in summer months [29, 30, 31]. If the tendencies continue in the near future, warmer summers and a bit cooler winters are to be expected in the city. These differentiated warming patterns may have important impacts on energy consumption, water supply, human health and natural environment in Istanbul. Therefore, future research should concentrate on the effects of climate change on different physical and human characteristics of the city. Apart from regular monitoring of climatic data, strategies and plans have to be put forward in order to cope with the negative effects of climate change. REFERENCES 1. 2. 3. 1672 Cazenave, A., C. Cabanes, K. Dominh, M.C. Gennero and C. Le Provost, 2003 Present-day sea level change: Observations and causes, Space Science Reviews, 108: 131-144. Bosello, F., R. Roson and R.S.J. Tol, 2007. Economwide estimates of the implications of climate change: Sea level rise, Environmental and Resource Economics, 37: 549-571. Bolch, T., 2007. Climate change and glacier retreat in northern Tien Shan (Kazakhstan/Kyrgyzstan) using remote sensing data, Global and Planetary Change, 56: 1-12. World Appl. Sci. J., 12 (10): 1662-1675, 2011 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Masiokas, M., R. Villalba, B.H. Luckman, M.E. Lascano, S. Delgado and P. Stepanek, 2008. 20th century glacier recession and regional hydroclimatic changes in northwestern Patagonia, Global and Planetary Change, 60: 85-100. Epstein, P.R., 2005. Climate change and human health, The new England J. Med., 353: 1433-1436. Khasnis, A. and M. Nettleman, 2005. Global warming and infectious disease, Archives of Medical Research, 36: 689-696. Emanuel, K., 2005. Increasing destructiveness of tropical cyclones over the past 30 years, Nature, 436: 686-688. Ramon, M.M. and J. Schwartz, 2007. Temperature, temperature rextremes and mortality: a study of acclimatisation and effect modification in 50 US cities, Occupational and Environ. Med., 64: 827-833. IPCC., 2007. Summary for Policymakers. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jones, P.D. and A. Moberg, 2003. Hemispheric and large-scale surface air temperature variations: An extensive revision and an update to 2001, Journal of Climate, 16: 206-223. Hanseni, J., M. Sato, R. Ruedy, K. Lo, D.W. Lea and M. Medina-Elizade, 2006. Global temperature change, Proceedings of the National Academy of Sciences of the United States of America, 103: 14288-14293. Rebetez, M. and M. Reinhard 2008. Monthly air temperature trends in Switzerland 1901-2000 and 1975-2004, Theoretical and Appl. Climatol., 91: 27-34. Stafford, J.M., G. Wendler and J. Curtis, 2000. Temperature and precipitation of Alaska: 50 year trend analysis, Theoretical and Appl. Climatol., 67: 33-44. Luterbacher, J., D. Dietrich, E. Xoplaki, M. Grosjean and H. Wanner, 2004. European seasonal and annual temperature variability, trends and extremes since 1500, Science, 303: 1499-1503. Founda, D., K.H. Papadopoulos, M. Petrakis, C. Giannakopoulos and P. Good, 2004. Analysis of mean, maximum and minimum temperature in Athens from 1897 to 2001 with emphasis on the last decade: trends, warm events and cold events, Global and Planetary Change, 44: 27-38. 16. Boo, K.O., W.T. Kwon and H.J. Baek, 2006. Change of extreme events of temperature and precipitation over Korea using regional projection of future climate change. Geophys Res 33, L01701, doi:10.1029/ 2005GL023378. 17. Brunet, M., P.D. Jones, J. Sigro, O. Saladie, E. Aguilar, A. Moberg, P.M. Della-Marta, D. Lister, A. Walther and D. Lo´pez, 2007. Temporal and spatial temperature variability and change over Spain during 1850-2005, J. Geophys. Res., 112, D12117, doi:10.1029/ 2006JD008249. 18. Beniston, M., D.B. Stephenson, O.B. Christensen, C.A.T. Ferro, C. Frei, S. Goyette, K. Halsnaes, T. Holt, K. Jylha, B. Koffi, J. Palutikof, R. Scholl, T. Semmler and K. Woth, 2007. Future extreme events in European climate: An exploration of regional climate model projections, Climatic Change, 81: 71-95. 19. Przybylak, R., 2007. Recent air-temperature changes in the Arctic, Ann. Glaciol., 46: 316-324. 20. Alexander, L.V., P. Hope, D. Collins and B. Trewin, 2007. Trends in Australia’s climate means and extremes: A global context, Aust. Met. Mag., 56: 1-18. 21. Toreti, A. and F. Desiato, 2008. Temperature trend over Italy from 1961 to 2004, Theoretical and Appl. Climatol., 91: 51-58. 22. Zhang, X., L.A. Vincent, W.D. Hogg and A. Niitsoo, 2000. Temperature and precipitation trends in Canada during the 20th century, Atmosphere-Ocean, 38: 395-429. 23. Pielke, R.A., C.A. Davey, D. Niyogi, S. Fall, J. Steinweg-Woods, K. Hubbard, X. Lin, M. Cai, Y.K. Lim, H. Li, J. Nielsen-Cammon, K. Gallo, R. Hale, R. Mahmood, S. Foster, R.T. McNider and P. Blanken, 2007. Unresolved issues with the assessment of multidecadal global land surface temperature trends, J. Geophysical Res., 112, D24S08,doi:10. 1029/2006JD008229. 24. Caesar, J., L. Alexander and R. Vose, 2006. Largescale changes in observed daily maximum and minimum temperatures: Creation and analysis of a new gridded data set, J. Geophysical Res., 111, D05101, doi:10.1029/2005JD006280. 25. Miro, J.J., M.J. Estrela and M. Milla´n, 2006. Summer temperature trends in a Mediterranean area (Valencia Region), Intl. J. Climatol., 26: 1051-1073. 26. Turkes, M., U.M. Sumer and G. Kilic, 1995. Variations and trends in annual mean air temperatures in Turkey with respect to climatic variability. Intl. J. Climatol., 15: 557-569. 1673 World Appl. Sci. J., 12 (10): 1662-1675, 2011 27. Alexandrow, V., S. Dakova, H. Aksoy and A. Dahamsheh, 2004. Analysis of climate change in southeastern Bulgaria and northwestern Turkey, Proceedings of the Conference on Water Observation and Information System for Decision Support, pp: 1-11, BALWOIS, Ohrid, FY Republic of Macedonia, 25-29 May 2004, http://balwois.mpl.ird.fr/ balwois/administration/full_paper/ffp-1o-148.pdf Accessed 16 June 2009. 28. Sönmez, F.K., A.Ü. Kömü çü, A. Erkan and E. Turgut, 2005. An analysis of spatial and temporal dimension of drought vulnerability in Turkey using the standardized precipitation index, Natural Hazards, 35: 243-264. 29. Partal, T. and E. Kahya, 2006. Trend analysis in Turkish precipitation data, Hydorological Processes, 20: 2011-2026. 30. Tecer, L.H., 2009. Temperature trends and changes in Rize, Turkey for the period 1975 to 2007, Clean, 37(2): 150-159. 31. Tayanç, M., U. Im, M. Do ruel and M. Karaca, 2009. Climate change in Turkey for the last half century, Climatic Change, 94: 483-502 32. Albek, M., 2008. The effects of climate change on rivers in the mediterreanean region of Turkey, Fresenius Environmental Bulletin, 17(9b): 1492-1500 33. Tatl , H., H.N. Dalfes and S.S. Mente , 2005. Surface air temperature variability over Turkey and its connection to large-scale upper air circulation via multivariate techniques. Intl. J. Climatol., 25: 331-350. 34. Price, D.T., D.W. McKenney, I.A. Nalder, M.F. Hutchinson and J.L. Kesteven, 2000. A comparison of two statistical methods for spatial interpolation of Canadian monthly climate data. Agr. For. Meteorol., 101: 81-94. 35. Rodriquex-Lado, L., G. Sparovek, P. Vidal-Torrado, D. Dourado-Neto and F. Macias-Vazquez, 2007. Modelling air temperature for the state of Sao Paulo, Brazil, Scientia Agricola, 64, doi: 10.1590/S010390162007000500002. 36. Thiessen, A.H., 1911. Precipitation averages for large areas. Mon. Wea. Rev., 39: 1082-1084. 37. Peck, E.L. and M.J. Brown, 1962. An approach to the development of isohyetal maps for moun tainous areas. J. Geophys. Res., 67: 681-694. 38. Schermerhorn, V.P., 1967. Relations between topography and annual precipitation in western Oregon and Washington. Water Res. Res., 3: 707-711. 39. Hutchinson, M.F. and R.J. Bischof, 1983. A new method of estimating mean seasonal and annual rainfall for the Hunter Valley, New South Wales. Aust. Meteorol. Mag., 31: 179-184. 40. Dingman, S.L., D.M. Seely-Reynolds and R.C. Reynolds, 1988. Application of kriging to estimating mean annual precipitation in a region of orographic influence. Water Resources B., 24: 329-339. 41. Phillips, D.L., J. Dolph and D. Marks, 1992. A comparison of geostatistical procedures for spatial analysis of precipitation in mountainous terrain. Agric. For. Meteorol., 58: 119-141 42. Daly, C., R.P. Nielson and D.L. Phillips, 1994. A statistical-topographic model for mapping climatological precipitation over mountainous terrain. J. Appl. Meteorol., 33: 140-158. 43. Nalder, I.A. and R.W. Wein, 1998. Spatial interpolation of climatic normals: test of a new method in the Canadian boreal forest. Agr. For. Meteorol., 9: 211-225 44. Reed, W.G. and J.B. Kince, 1917. The preparation of precipitation charts. Mon. Wea. Rev., 45: 233-235. 45. Spreen, W.C., 1947. A determination of the effect of topography upon precipitation. Trans. Am. Geophys. Union, 28: 285-290. 46. Burns, J.I., 1953. Small-scale topographic effects on precipitation distribution in San Dimas Experimental Forest. Trans. Am. Geophys. Union, 34: 761-768. 47. Houghton, J.G., 1979. A model for orographic precipitation in the North-Central Great Basin. Mon. Wea. Rev., 107: 1462-1475. 48. Gandin, L.S., 1963. Objective Analysis of Meteorological Fields. Gidrometeoizdat, Leningrad, p. 287. (English translation: Israel Program for Scientific Translations, Jerusalem, 1965. 49. Bayraktar, H.F., F.S. Turalioglu and Z. Sen, 2005. The estimation of average areal rainfall by percentage weighting polygon method in Southeastern Anatolia Region, Turkey. Atmospheric Res., 73(1-2): 149-160. 50. Cheung, W., G.B. Senay and A. Singh, 2008. Trends and spatial distribution of annual and seasonal rainfall in Ethiopia. Intl. J. Climatol., 28(13): 1723-1734. 51. Chaoka, R.T., B.F. Alemaw and D.M. Tsiege, 2007. Modeling and Understanding the Relationship between Vegetation and Rainfall of Tropical Watershed using Remote Sensing Data and GIS. J. Spatial Hydrol., 7(2): 47-61. 1674 World Appl. Sci. J., 12 (10): 1662-1675, 2011 52. Fiedler, F.R., 2003. Simple, Practical Method for Determining Station Weights Using Thiessen Polygons and Isohyetal Maps. J. Hydrol. Engineering, 8(4): 214-218. 53. Li, M.H., M.J. Yang, R. Soong and H.L. Huang, 2005. Simulating Typhoon Floods with Gauge Data and Mesoscale-Modeled Rainfall in a Mountainous Watershed J. Hydrometeorol., 6: 306-323. 54. Tumas, 2009 Meteorolojik Veri Ar iv Sistemi http://tumas.dmi.gov.tr/wps/portal. Accessed 20 June 2009. 55. Tonkaz, T., M. Çetin and K. Tülücü, 2007. The impact of water resources development projects on water vapor pressure trends in a semi-arid region, Turkey Climatic Change, 82:195-209. 56. Türke , M., 2002. Spatial and Temporal Variations in Precipitation and Aridity Index Series of Turkey. in: Mediterranean Climate Variability and Trends. Regional Climate Studies. Springer Verlag, Heidelberg, pp: 181-213. 57. Sneyers, R., 1990. On the statistical analysis of series of observations. WMO, Technical Note No. 143, Geneva, Switzerland, pp: 192. 58. Kukul, Y.S., S. Anaç, E. Yeþil rmak and J.M. Moreas, 2007. Trends of precipitation and stream-flow in Gediz River Basin, Western Turkey, Fresenius Environmental Bulletin, 16(5): 477-488. 59. Gan, Y.T., 1998. Hydroclimatic Trends and Possible Climatic Warming in the Canadian Praires. Water Resources, 34(11): 3009-3015. 60. Yue, S. and M. Hashino, 2003. Long Term Trends of Annual and Monthly Precipitation in Japan. J. American Water Resources Association, 39(3): 587-596. 61. Fu, G., S. Chen, C. Liu and D. Shepard, 2004. HydroClimatic Trends of the Yellow River Basin for the Last 50 Years. Climatic Change, 65(1-2): 149-178. 62. Gemmer, M., S. Becker and T. Jiang, 2004. Observed Monthly Precipitation Trends in China 1951-2002. Theoretical and Appl. Climatol., 77(1-2): 39-45. 63. Wang, W., X. Chen, P. Shi and P.H.A.J.M. van Gelder, 2008. Detecting changes in extreme precipitation and extreme streamflow in the Dongjiang River Basin in southern China. Hydrol. Earth Syst. Sci., 12: 207-221. 64. TÜIK., 2008. Adrese dayal nüfus kay t sistemi (ADNKS), 2007 Nüfus Say m Sonuçlar . http://tuikapp.tuik.gov.tr/adnksdagitimapp/adnks.zu. Accessed 29 July 2008. 65. Tayanç, M., 2000. An assessment of spatial and temporal variation of sulfur dioxide levels over Istanbul, Turkey, Environ. Pollution, 101: 61-69. 66. Anteplio lu, U., S. Topçu and S. Incecik, 2002. Air Pollution Modeling and Its Application XV, Edited by Borrego and Schayes “An application of a photochemical model for urban airshed in Istanbul, Kluwer Academic/Plenum Publishers, New York, pp: 167-175. 67. Demirci, A., 2001. Types and distribution of landslides in the eastern parts of Buyukcekmece Lake, using GIS, Unpublished Master Thesis, Fatih University, Social Science Institute, Istanbul. 68. Karaburun, A., A. Demirci and Suen I-Shian, 2009. Impacts of urban growth on forest cover in Istanbul (1987-2007), Environmental Monitoring and Assessment, DOI 10.1007/s10661-009-1000-z 69. Gold, C.M., 1991. Problems with handling spatial data--the Voronoi approach. C.I.S.M. J., 45: 65-80. 70. Ball, J.E. and K.C. Luk, 1998. Modeling Spatial Variability of Rainfall Over A Catchment, J. Hydrol. Engineering, 3: 122-130. 71. Yu, Y.S., S. Zou and D. Whittemore, 1993. Nonparametric trend analysis of water quality data of rivers in Kansas. J. Hydrol., 150: 61-80. 72. Sen, P.K., 1968. Estimates of regression coefficients based on Kendall’s tau. Journal of the American Statistical Association, 63(1968): 1379-1389. 1675