Atmospheric Research 138 (2014) 73–90

Contents lists available at ScienceDirect

Atmospheric Research

j ourna l homepage: www.e lsev ie r .com/ locate /atmos

Spatial and temporal trends of mean and extreme rainfall andtemperature for the 33 urban centers of the arid and semi-aridstate of Rajasthan, India

Santosh M. Pingale a,⁎, Deepak Khare a, Mahesh K. Jat b, Jan Adamowski c

a Department of Water Resource Development and Management, Indian Institute of Technology Roorkee, Roorkee 247 667 UK, Indiab Department of Civil Engineering, Malaviya National Institute of Technology, Jaipur, Rajasthan, Indiac Department of Bioresource Engineering, McGill University, Quebec, Canada

a r t i c l e i n f o

⁎ Corresponding author currrently at: Department oIrrigation Engineering, Arba Minch University, Ethiopi

E-mail address: pingalesm@gmail.com (S.M. Pinga

0169-8095/$ – see front matter © 2013 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.atmosres.2013.10.024

a b s t r a c t

Article history:Received 7 May 2013Received in revised form 21 September 2013Accepted 29 October 2013

Trend analysis of the mean (monsoon season, non-monsoon season and annual) and extremeannual daily rainfall and temperature at the spatial and temporal scales was carried out for all the33 urban centers of the arid and semi-arid state of Rajasthan, India. Statistical trend analysistechniques, namely theMann–Kendall test andSen's slope estimator,were used to examine trends(1971–2005) at the 10% level of significance. Both positive and negative trends were observed inmean and extreme events of rainfall and temperature in the urban centers of Rajasthan State. Themagnitude of the significant trend of monsoon rainfall varied from (−) 6.00 mm/hydrologic yearat Nagaur to (−) 8.56 mm/hydrologic year at Tonk. However, the magnitude of the significantnegative trends of non-monsoon rainfall varied from (−) 0.66 mm/hydrologic year at Dungarpurto (−) 1.27 mm/hydrologic year at Chittorgarh. Themagnitude of positive trends of non-monsoonrainfall varied from 0.93 mm/hydrologic year at Churu to 1.70 mm/hydrologic year atHanumangarh. The magnitude of the significant negative trends of annual rainfall varied from(−) 6.47 mm/year at Nagaur to (−) 10.0 mm/year at Tonk. Theminimum, average andmaximumtemperature showed significant increasing warming trends on an annual and seasonal scale inmost of the urban centers in Rajasthan State. The magnitude of statistically significant annualextreme daily rainfall varied from 2.00 mm at Jhalawar to (−) 1.64 mm at Tonk, while themagnitude of statistically significant extreme annual daily minimum and maximum temperaturevaried from 0.03 °C at Ganganagar to 0.05 °C at Jhalawar, respectively. The spatial variations of thetrends in mean (monsoon season, non-monsoon season and annual) and extreme annual dailyrainfall and temperature were also determined using the inverse-distance-weighted (IDW)interpolation technique. IDW results are helpful to identify trends and variability in mean andextreme rainfall and temperature in space and time for the study locations where the data is notavailable and the quality of data is not good. These spatial maps of temperature and rainfall canhelp local stakeholders and water managers to understand the risks and vulnerabilities related toclimate change in terms of mean and extreme events in the region.

© 2013 Elsevier B.V. All rights reserved.

Keywords:Extreme eventsInterpolationMann–Kendall testTemperatureRainfallIndia

1. Introduction

Increasing urbanization, populations and economies arecausing adverse impacts on the natural environment. With

f Water Resources anda.le).

ll rights reserved.

increases in urbanization, problems related to climate changeare becoming worse, since additional pervious areas arebecoming impervious, leading to decreases in vegetation coverand increases in temperature. Rapid economic developmentand population growth in various parts of the world such asIndia have caused concerns regarding the quantity and qualityof natural resources, and in particular water resources. As such,information on the magnitude, duration and frequency of

74 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

extreme events (such as floods and droughts), along withtrends in various hydro-meteorological variables, is requiredfor the planning of optimum adaptation strategies for thesustainable use of natural resources such as water (e.g.,Kampata et al., 2008; Some'e et al., 2012; Huang et al., 2013).Mall et al. (2006) have studied the potential for sustainabledevelopment of surface water and groundwater resourceswithin the constraints imposed by climate change. The authorssuggested that little work has been done on hydrologicalimpacts of climate change for Indian regions/basins. Therefore,assessment of climate change is required in India whichconsiders the causative factors of climate change (i.e., anthro-pogenic factors, global warming in terms of variations inmeteorological parameters) at spatial and temporal scales.Further, it is essential in hydrological modeling to assess theadverse impacts of climate change on water resources. Suchstudies are even more significant for urbanized watersheds/areas to better understand the relationship between climatechange and hydrological processes, which will facilitate thesustainable utilization of water resources.

There are different methods that can be used in theassessment of climate change. Various parametric andnon-parametric statistical tests have been used in the recentpast by many researchers to assess trends in hydro-meteorological time series in India (e.g., Gadgil andDhorde, 2005; Singh et al., 2008a, 2008b; Basistha et al.,2009; Kumar et al., 2010; Pal and Al-Tabbaa, 2010; Deka etal., 2012; Duhan et al., 2013; Jeganathan and Andimuthu,2013; Jain et al., 2013), and in other regions of the world(e.g., Zhang et al., 2000; Zhang et al., 2001; Roy and Balling,2004; Cheung et al., 2008; Motiee and McBean, 2009; Sahooand Smith, 2009; Zin et al., 2010; Tabari et al., 2011; Nalley etal., 2012; Nemec et al., 2012; Saboohi et al., 2012; Jiang et al.,2013). Detection of past trends, changes, and variability in thetime series of hydro-climatic variables is very important inunderstanding the potential impact of future changes in theregion (Sahoo and Smith, 2009). Statistical analysis could beextended to analyze climatic parameters and their relationshipswith water resources, land use/cover changes, urbanization etc.

It has been seen that the Earth's average temperature hasincreased by 0.6 °C in the later part of the 20th century; thereis also a dramatic shift in temperature change from aminimum of 1.4 °C to a maximum of 5.4 °C as reported bythe projections made by various climate prediction models(IPCC, 2001). The IPCC has demonstrated that climate changeassessments at sub-national (i.e., state or province) and localscales are needed. The IPCC has also found several gaps in theknowledge that exists in terms of observations and researchneeds related to climate change and water (IPCC, 2007).These include GCM scale resolutions and downscalingtechniques to locale scales to assess the climate change andto quantify the contribution of anthropogenic activity toclimate change etc.

It has been found that seasonal and annual air temperatureshave been increasing at a rate of 0.57 °C per 100 years (1881 to1997) in India (Pant and Kumar, 1997). The trends of certainhydro-climatic variables have been studied in Indian urbancenters such asHyderabad, Patna, Ahmedabad, Surat, Bangalore,Mumbai, Nagpur, Pune, Jaipur, Chennai, New Delhi, Kanpur,Lucknow and Kolkata, whose populations are more than1 million (De and Rao, 2004). Significant increasing trends

were found in annual and monsoon rainfall over Chennai, NewDelhi, Kolkata andMumbai. Other climatic parameters were notconsidered to assess climate change impacts on timing,distribution, pattern, frequency, variability of precipitation andtheir inter-relationships in urban cities. Arora et al. (2005)explored trends of annual average and seasonal temperaturesusing the MK test at the country and regional scales in India. Itwas observed that annual mean temperature, mean maximumtemperature and mean minimum temperature have increasedat the rates of 0.42, 0.92 and 0.09 °C (100 year)−1, respectively.On a regional basis, stations in Southern and Western Indiashow rising trends of 1.06 and 0.36 °C (100 year)−1, respec-tively, while stations in the North Indian plains show a fallingtrend of 0.38 °C (100 year)−1. The seasonal mean temperaturehas increased by 0.94 °C (100 year)−1 for the post-monsoonseason and by 1.1 °C (100 year)−1 for the winter season. Thisstudy deals only with trend detection of temperature; inaddition, trends in annual extreme daily events of rainfall andtemperature have not been studied.

Basistha et al. (2007) assessed the spatial trends of rainfallover Indian subdivisions from 1872 to 2005. Their results showdecreasing trends of rainfall over North India excluding Punjab,Haryana, West Rajasthan, and Saurashtra, and increased trendsin South India excluding Kerala and Madhya Maharashtra. Thestudy also concluded that the arid portion (i.e., western part)of India has not been investigated in detail. Further, theyhighlighted that more research is required to assess the spatialpatterns of trends of other climatic variables (average, mini-mum and maximum temperatures, relative humidity, windspeed, evapotranspiration (ET), number of rainy days etc.) andtheir inter-relationships including trends of annual and seasonalclimatic parameters to assess the impacts of climate change.Gowda et al. (2008) studied the region of Davangere districtover a period of 32 years using statistical analysis. Climaticparameters (i.e., rainfall, relative humidity, maximum temper-ature, minimum temperature, sunshine hour and wind speed)were analyzed to assess climate change. The statistical analysisshowed that such a small data set may not represent the correctpicture of climate change and requires long term data. Asignificant change in climatic variables was found in and aroundthe Davangere region. Ghosh et al. (2009) observed the variedtrend in Indian summer monsoon rainfall, which has not onlybeen affected by global warming, but may also be affected bylocal changes due to rapid urbanization, industrialization anddeforestation. Patra et al. (2012) detected rainfall trends inthe twentieth century in India using parametric and non-parametric statistical trend analysis tests. The temporal varia-tion inmonthly, seasonal and annual rainfall was studied for theOrissa State using data from 1871 to 2006. The analysis revealeda long term, insignificant, declining trend in annual as well asmonsoon rainfall, and an increasing trend in the post-monsoonseason. Rainfall during winter and summer seasons showed anincreasing trend. The study only assessed rainfall for OrissaState; trends inmean and extreme annual daily eventswere notexplored. Duhan andPandey (2013) investigated the spatial andtemporal variabilities of precipitation in 45 districts of MadhyaPradesh (MP), India on an annual and seasonal basis. The MKtest and Sen's slope estimator test were used for trend analysis,and an increase and decrease in the precipitation trend werefound on annual and seasonal bases, respectively, in the districtsof MP. In this study, little attention was given to urban trends of

75S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

mean and extreme rainfall and temperature. A literature reviewshows that in India climate related studies are restricted tostatistical analysis of some of the meteorological parametersonly. The effects of climate change on water resources were notconsidered and seasonal changes are not clearly indicated withmean and extreme events, especially in the urban centers.

As described above, the analysis of extreme trend andvariability in annual daily rainfall and temperature has beencarried out in recent years by a number of researchers in variousparts of theworld, including India. However, there is a still a lackof information on trends and variability of extreme annual dailyevents, especially for the semi-arid and arid region of Rajasthan,India. For the region of Rajasthan, there is a lack of studiesof climate change assessments that consider various regionspecific aspects, and their inter-relation with the climaticparameters at the spatial and temporal scales. Climate changein terms of inter-seasonal and inter-annual variations in meanand extreme annual daily rainfall and temperature, for urbancenters of arid and semi-arid regions in India, has not beenreported in much detail in the literature. In light of this, thepresent study was aimed at studying climate change in terms ofmean (monsoon, non-monsoon and annual) and extremeannual daily events in the 33 main urban centers of Rajasthanin India via trend analysis of two important climatic parameters(i.e., rainfall and temperature (minimum, average andmaximum)) at the spatial and temporal scales. Statistical trendanalysis techniques, namely the Mann–Kendall (MK) test(Mann, 1945; Kendall, 1975) and Sen's slope estimator (Sen,1968), were used in the present study. These methods haveseveral advantages over parametric methods (Duhan andPandey, 2013).

The paper is organized in five sections. Section 2 describesthe study area and data used. Section three describes themethodology adopted in the study. Sections 4 and5 describe thefindings of the study on climate change assessment in terms oftrends in mean (monsoon season, non-monsoon season andannual) and extreme annual daily rainfall and temperature inthe 33 urban centers of Rajasthan State in India. Further, thespatial variations in mean (seasonal and annual) and extremeannual daily rainfall and temperature with their trends arepresented using the inverse-distance-weighted (IDW) interpo-lation method.

2. Study area and data used

Rajasthan is the largest state in India, with an area of342,000 km2 of which about 60% falls under the arid category.Administratively, the state is divided into 33 districts (Fig. 1)and its climate varies from arid to semi-arid. Geographically,Rajasthan State extends from 23° 3.50′ to 30° 14′N latitude and69° 27′ to 78° 19′E longitude. The study region has uniqueclimatological characteristics (low and erratic rainfall, extremesof diurnal and annual temperatures, low humidity and highwind velocity). It also has unique physiographic characteristics.For example, the Arawali Hill Range is unique because it liesalong the direction of the southwest monsoon, which bringsrains in the region. There is also the Great Indian Desert (calledthe Thar Desert) in the northwestern part of Rajasthan, which isIndia's largest desert, covering almost 70% of the state; in thisregion there is less vegetation and the region has a significantinfluence on rainfall and temperature distribution across the

state. The climate in the state varies from arid to semi-arid. Thismakes Rajasthan State different from other regions in neigh-boring states. The state is bordered by Gujarat, Madhya Pradesh,Uttar Pradesh, Haryana and Punjab, and is also bordered withPakistan. The climate in Rajasthan State is characterized by lowand erratic rainfall, extremes of diurnal and annual tempera-tures, low humidity and high wind velocity. The annual averagerainfall is highly variable, and it is most erratic in the westernregion with frequent dry spells, punctuated occasionally byheavy downpours in some years associated with passing lowpressure systems over the region. The western and southerndistricts of Rajasthan State frequently experience severedroughts. Coping with low rainfall, with a high coefficient ofvariation across time and space, is the major challenge (Rathor,2005).

Gridded data sets of rainfall and temperature have beenused in many hydrological and climatological studies world-wide, including Australia, for hydro-climatic forecasting,climate attribution studies and climate model performanceassessments (e.g., Tozer et al., 2012; Dash et al., 2013).Therefore, in the present study 0.5° × 0.5° and 1° × 1°gridded data sets of daily rainfall and temperature from theperiod 1971 to 2005 were used for the spatial and temporalvariability and trend analyses in the present study. Thesedata were obtained from the Indian Meteorological Depart-ment (IMD), Pune. From all of India, 395 stations were usedfor developing the high resolution daily gridded data sets(1° × 1°) of daily temperature (minimum, maximum andmean) for the Indian region by the Indian MeteorologicalDepartment (IMD) — Pune. The modified Shepard's angulardistance weighting algorithm was used for interpolating thestations of temperature data into 1° × 1° grids. These datasets were evaluated using the cross validation technique toestimate the interpolation error (root mean square error),which was found to be less than 0.5 °C. These data sets werealso compared with another (Asian Precipitation — HighlyResolved Observational Data Integration Towards the Evalu-ation of Water Resources (APHRODITE)) high resolutionmonthly data set, and a correlation of more than 0.8 wasfound in most parts of the country. Similarly, high resolutiondaily gridded data sets of rainfall (0.5° × 0.5°) were alsodeveloped (using Shepard's method), and validated using thenetwork of 6076 rain gage stations for the Indian region byIMD, Pune. Additional detail on the methodology is given inIMD reports (Rajeevan and Bhate, 2008; Srivastava et al.,2008).

These data sets have been found to be useful for variousregional applications (such as studies of extreme tempera-tures, validation of numerical and climate model simulationsand many environmental applications including trend anal-ysis of mean and extreme events of rainfall and temperature)(Rajeevan and Bhate, 2008 and Srivastava et al., 2008), andthey can be used for water resource planning purposes. Thedata was subjected to thorough quality checks (location ofstations, typing error, missing data etc.) before the dataanalysis was performed. Hamlet et al. (2005) have reportedthat existing gridded data sets are not appropriate for thelong term trend estimation of precipitation and temperaturefields for the continental United States. The griddingprocedures introduce artificial trends due to the incorpora-tion of stations with different record lengths and locations.

Fig. 1. Location of urban centers considered in the state of Rajasthan, India.

76 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

However, the gridded set developed by IMD, Pune addressedthese issues before and after developing the gridded data setsfrom the observed station data before releasing the datato users. In addition, these data sets were validated withother gridded data sets developed in the region, and theirapplication was demonstrated successfully through casestudies and by other researchers (e.g., Dash et al., 2013).

Basic data quality checks (such as rejecting values, greaterthan exceeding known extreme values, minimum tempera-ture, greater than maximum temperature, same temperaturevalues for many consecutive days, unusual high values,

homogeneity, location of the station etc.) were performedfor each station, and then stations were selected based ondata quality or the development of the gridded data sets oftemperature and rainfall. Further, homogeneity, persistenceand periodicity in the data were also analyzed by autocorre-lation. Normalizing the time series before the analysis wasperformed at each urban center. The district headquarters ofRajasthan State, which has the largest population, greatestdevelopmental activities, and administrative centers of thedistricts, are considered as the ‘urban centers’. The geograph-ical extent of the urban centers is defined by their municipal

77S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

boundaries defined by the development authorities. The sizesand populations of the urban centers are included in Table 1.The climate of individual urban centers has been extractedfrom the gridded surface of temperature and rainfallcorresponding to the municipal limits of the urban centers.Therefore, it preserves the meteorological characteristics ofindividual urban centers. The IDW assumes that the values ofthe variables at a point to be predicted are similar to thevalues of nearby observation points. The majority of thestations are located in the urban areas. Therefore, this has amajor influence of urban climate. The grid covering the urbancenter of the district headquarters of Rajasthan State isconsidered to represent the climate of the urban center. First,the grids that represent the urban center from the developeddata sets are identified, and then daily data (1971–2005) wasextracted. The daily rainfall and temperature (minimum,average and maximum) data were extracted for the 33 urbancenters of district headquarters in Rajasthan State (Table 1).These urban centers were selected taking into account thelength and availability of records, so that most of the urbancenters in the arid and semi-arid region of Rajasthan State werecovered by the corresponding data to assess trends in mean(seasonal and annual) and extreme annual daily rainfall andtemperature. The grids covering the urban centers were chosen

Table 1List of urban centers, size and population.

ID Urban center Longitude Latitude Elev

1 Ajmer 74° 42′ E 26° 27′ N 4862 Alwar 76° 38′ E 27° 34′ N 2713 Banaswara 74° 24′ E 23° 30′ N 3024 Baran 76° 33′ E 25° 05′ N 2625 Barmerb 71° 25′ E 25° 45′ N 2276 Bharatpur 77° 30′ E 27° 15′ N 1837 Bhilwara 74° 40′ E 25° 21′ N 4218 Bikaner 73° 22′ E 28° 01′ N 2429 Bundi 75° 41′ E 25° 27′ N 26810 Chittorgarh 74° 42′ E 24° 54′ N 39411 Churu 75° 01′ E 28° 19′ N 29212 Dausab 76° 20′ E 26° 53′ N 33313 Dholpur 77° 52′ E 26° 41′ N 17714 Dungarpur 73° 43′ E 23° 50′ N 22515 Ganganagar 73° 50′ E 29° 49′ N 16416 Hanumangarh 74° 21′ E 29° 35′ N 17717 Jaipur 75° 52′ E 26° 55′ N 43118 Jaisalmer 70° 57′ E 26° 55′ N 22919 Jalore 72° 58′ E 25° 22′ N 17820 Jhalawarb 76° 10′ E 24° 35′ N 31221 Jodhpur 73° 04′ E 26° 18′ N 23122 Jhunjhunu 75° 20′ E 28° 06′ N 32323 Karauli 77° 04′ E 26° 30′ N 27524 Kota 75° 52′ E 25° 10′ N 27125 Nagaur 73° 40′ E 27° 00′ N 30226 Pali 73° 25′ E 25° 46′ N 21427 Rajsamandb 73° 53′ E 25° 04′ N 54728 Sawai Madhopur 76° 21′ E 26° 01′ N 26629 Sikar 75° 15′ E 27° 36′ N 42730 Sirohib 72° 51′ E 24° 53′ N 32131 Tonk 75° 47′ E 26° 10′ N 28932 Udaipur 73° 42′ E 24° 34′ N 59833 Pratapgarh 74° 47′ E 24° 02′ N 491

a For size of urban center of district headquarter Rajasthan see e.g., Jhala0x39653ede53cfaf07:0x31263af08bba421e,Jhalawar,+Rajasthan&gl=in&ei=TekFUhttp://www.citypopulation.de/php/india-rajasthan.php.

b Population census 2001.

for extracting the daily rainfall and temperature data, and it wasassumed that the grids covering the urban center represent therainfall and temperature of that urban center in Rajasthan Stateover a period of time (Fig. 1 and Table 1). Further, the rainfalland temperature (minimum, average andmaximum) datawerearranged according to monsoon season, non-monsoon seasonand annual scale from 1971 to 2005 in each urban center. Themonsoon season represents the months of June to September,and the non-monsoon season represents the months of Januaryto May and October to December.

3. Methodology

3.1. Normalization and autocorrelation analysis of time series

Normalized rainfall and temperature time series wereused to test the outliers present in the time series data usingfollowing relationship (Rai et al., 2010):

Xt ¼ xt‐xð Þ=σ ð1Þ

where Xt is the normalized anomaly of the series, xt is theobserved time series, and x and σ are the long-term mean andstandard deviation of annual/seasonal time series, respectively.

ation (m.a.s.l.), m Size (km2)a

approx.Population (2011 census)

31 542,580– 315,310– 100,128– 118,157– 83,517– 252,109– 360,00965 647,804– 102,823– 116,409– 119,846– 61,589– 126,142– 43,000– 224,773– 151,104168 3,073,350– 78,000– 54,185– 48,05483 1,033,918– 118,966– 105,69062 1,001,365– 100,618– 229,956– 55,671– 120,998– 237,579– 35,544– 165,363– 451,735– 48,900

war, https://maps.google.co.in/maps?q=Jhalawar&ie=UTF8&hq=&hnear=uK8G4bLrQfk74F4&ved=0CLABELYD. More details are also available in

78 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

The autocorrelation test was performed to check therandomness and periodicity in the time series (Modarres andda Silva, 2007). If lag-1 serial coefficients are not statisticallysignificant then the MK test can be applied to the originaltime series (Luo et al., 2008; Karpouzos et al., 2010). Theserial correlation coefficients of the normalized climaticseries were computed for lags L = 0 to k, where k is themaximum lag (i.e., k = n/3), and n is the length of the series.The autocorrelation coefficient rk of a discrete time series forlag-k was estimated as:

rk ¼

Xn‐kk¼1

Xt−Xt� �

Xtþk−Xtþk� �

Xn‐kk¼1

Xt−Xt� �2 Xtþk−Xtþk

� �2" #0:5 ð2Þ

where rk is the lag-k serial correlation coefficient. The hypothesisof serial independence was then tested by lag-1 autocorrelationcoefficient as H0 : r1 = 0 against H1 : |r1| ≥ 0 using the test ofsignificance of serial correlation (Yevjevich, 1971; Rai et al.,2010):

rkð Þtg ¼−1� tg n‐k‐ð Þ0:5

n‐kð3Þ

where rkð Þtg is the normally distributed value of rk, and tg is thenormally distributed statistic at ‘g’ level of significance. Thevalues of tg are 1.645, 1.965 and 2.326 at the significance level of0.10, 0.05 and 0.01, respectively. The null hypothesis of serialindependence was rejected at the significance level α (here0.05), if rkj j≥ rkð Þtg . For the non-normal time series data, the MKtest is an appropriate choice for trend analysis (Yue and Pilon,2004). Therefore, the MK test was used in the present study,where autocorrelation was found to be non-significant at the 5%level of significance.

3.2. Mann–Kendall test

The Mann–Kendall (MK) test is a non-parametric test thatcan be used for detecting trends in a time series (Mann, 1945)where autocorrelation is non-significant. The non-linear trend,as well as the turning point can be derived from Kendall teststatistics (Kendall, 1975). This method searches for a trend in atime series without specifying whether the trend is linear ornon-linear. It has been found to be an excellent tool for trenddetection and many researchers have used this test to assessthe significance of trends in hydroclimatic time series data suchas water quality, stream flow, temperature and precipitation(e.g., Ludwig et al., 2004; Zhang et al., 2004;McBean andMotiee,2008; Basistha et al., 2009; Rai et al., 2010; Patra et al., 2012).TheMK test can be applied to a time series xi ranked from i = 1,2 … n − 1 and xj ranked from j = i + 1, 2 … n such that:

sgn xj−xi� �

¼1 if xj−xi

� �N0

0 if x j−xi� �

¼ 0

−1 if xj−xi� �

b0

:

8>>><>>>:

ð4Þ

The Kendall test statistic S can be computed as:

S ¼Xn−1

k¼1

sgn xj−xk� �

ð5Þ

where sgn(xj − xk) is the signum function. The test statistic Sis assumed to be asymptotically normal, with E(S) = 0 forthe sample size n ≥ 8 and variance as follows:

V Sð Þ ¼½n n−1ð Þ 2nþ 5ð Þ−

Xt

t t−1ð Þ 2tþ 5ð Þ�

18ð6Þ

where ti denotes number of ties up to sample i.The standardized MK test statistic (Zmk) can be estimated

as:

Zmk ¼

S−1ffiffiffiffiffiffiffiffiffiffiV Sð Þp if S N 0

0 if S ¼ 0Sþ 1ffiffiffiffiffiffiffiffiffiffiV Sð Þp if S b 0

:

8>>>><>>>>:

ð7Þ

The standardized MK test statistic (Zmk) follows thestandard normal distribution with a mean of zero andvariance of one. If ±Zmk ≤ Zα/2 (here α = 0.1), then thenull hypothesis for no trend is accepted in a two sided testfor trend, and the null hypothesis for the no trend is rejectedif ±Zmk ≥ Zα/2. Failing to reject H0 (i.e., null hypothesis) doesnot mean that there is no trend. Rather, it is a statement thatthe evidence available is not sufficient to conclude if there is atrend (Helsel and Hirsch, 2002). A positive value of Zmk

indicates an ‘upward trend’ and a negative value indicates a‘downward trend’. The significance levels (p-values) for eachtrend test can be obtained from the relationship given as(Coulibaly and Shi, 2005):

p ¼ 0:5−φ Zmkj jð Þ ð8Þ

where φ() denotes the cumulative distribution function (CDF)of a standard normal variate. At a significance level of 0.1, ifp ≤ 0.1 then the existing trend is considered to be statisticallysignificant.

3.3. Sen's estimator of slope and percentage change over a period

If a linear trend is present in a time series, then the trueslope of the trend is estimated using a simple non-parametricprocedure (Theil, 1950; Sen, 1968), which is given by:

Qi ¼ medianxj−xkj−k

� �∀k≤ j ð9Þ

where xj and xk are data values at times j and k (j N k),respectively. The median of N values of Qi is Sen's estimator ofslope. If N is odd, then Sen's estimator is computed byQmed = Q(N + 1)/2 and if N is even, then Sen's estimator iscomputed by Qmed = [QN/2 + Q(N + 2)/2]/2. Finally, Qmed is

79S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

tested by a two-sided test at 100(1 ‐ α)% confidence interval,and the true slope is obtained by a non-parametric test.

The percentage change over a period of time can beestimated from Theil and Sen's median slope and mean byassuming the linear trend (Yue and Hashino, 2003; Basisthaet al., 2009) in the time series:

% change ¼ median slope� length of periodmean

� �: ð10Þ

3.4. Interpolation technique

In the present study, the spatial variation and trends of themean (monsoon, non-monsoon and annual) and extremeannual daily rainfall and temperature were determined usingthe inverse-distance-weighted (IDW) interpolation technique.The IDW method is a simple and effective interpolationtechnique that is based on the assumption that the values ofthe variables at a point to be predicted are similar to the valuesof nearby observation points. This method implies that eachstation has a local influence, which decreases with distance bymeans of the use of a power parameter (Isaaks and Srivastava,1989; Robinson and Metternicht, 2006). The general formula ofIDW (Yavuz and Erdogan, 2012) is as follows:

Z S0ð Þ ¼ mXNi¼1

λiZ Sið Þ ð11Þ

where

Z S0ð Þ is the value of prediction for location S0N is the number of measured sample points sur-

rounding the prediction locationλi is the weight assigned to each measured pointZ(Si) is the observed value at the location Si.

The weighting is a function of inverse distance. Theformula determining the weighting is (Yavuz and Erdogan,2012):

λi ¼d−pioXN

i¼1

d−pio

: ð12Þ

The IDW method requires the choice of a power parameterand a search radius. The power parameter (p) controls thesignificance of measured values on the interpolated value basedupon their distance from the output point (d) (Erdoğan, 2009).The choice of a relatively high power parameter ensures a highdegree of local influence and gives the output surface in moredetail (Yavuz and Erdogan, 2012).

4. Results and discussion

Detailed trend analysis of the mean (monsoon season,non-monsoon season and annual) and extreme annual dailyrainfall and temperature was carried out for the period of1971–2005 using theMann–Kendall test for all 33 urban centersof Rajasthan State. Similarly, Sen's slope and percentage change

in rainfall and temperature were also determined. The corre-sponding spatial and temporal distributions of the mean(monsoon season, non-monsoon season and annual) andextreme events of annual daily rainfall and temperature usingthe MK test statistics (Zmk) were determined by the IDWinterpolation technique. The increasing, decreasing, and notrend at 10% level of significance are represented with thesymbols ↑, ↓ and О, respectively.

4.1. Trend analysis of mean events

4.1.1. RainfallThe spatial distributions of trends in seasonal (monsoon

and non-monsoon) and annual average rainfall for the 33urban centers of Rajasthan State are shown in Fig. 2. Themagnitude of the Sen's slope and percentage change inaverage rainfall at seasonal (monsoon and non-monsoon)and annual temporal scales for the 33 different urban centersof Rajasthan State is also given in Tables S1 to S3 (Supportingmaterial), respectively. Both positive and negative trendswere observed by the MK test and Sen's slope estimator inseasonal and annual average rainfall in the urban centers.This is in good agreement with observations in different partsof India (Pal and Al-Tabbaa, 2011).

All of the significant trends in monsoon and annual scalewere found to be decreasing at the 10% level of significance.However, both positive and negative trends in non-monsoonaverage rainfallwere observed. As a result, the expecteddecreasein rainfall will possibly cause a decrease in water availability inthe future (Gosain et al., 2006; Tabari and Talaee, 2011b). Themonsoon season rainfall pattern and trends were very similar tothe annual ones, which demonstrate the major contribution ofmonsoon rainfall in annual rainfall in the urban centers ofRajasthan State. Similar results indicating themajor contributionof seasonal precipitation in annual precipitation have alsobeen observed in Iran (Tabari and Talaee, 2011b). Significantdecreasing trends were observed in monsoon average rainfallfor Jodhpur (at −6.82 mm/hydrologic year and −88.16%),Karauli (at −7.59 mm/hydrologic year and −39.03%), Nagaur(at −6.00 mm/hydrologic year and −66.92%) and Tonk (at−8.56 mm/hydrologic year and −50.63%). The significantdecreasing trends of average monsoon rainfall had the highestvalue of decreasing slope (i.e., −8.56 mm/hydrologic year) forTonk, and percentage change of−88.16% in Jodhpur (Table S1).Significant trends in non-monsoon average rainfall were foundfor Dungarpur (at −0.66 mm/hydrologic year and −58.10%),Chittorgarh (at −1.27 mm/hydrologic year and −80.29%),Jaipur (at 1.22 mm/hydrologic year and 61.53%), Churu(at 0.93 mm/hydrologic year and 49.61%), Ganganagar (at1.39 mm/hydrologic year and 79.46%) and Hanumangarh (at1.70 mm/hydrologic year and 102.85%). A significant increasingtrend was observed in non-monsoon rainfall at Hanumangarh(1.70 mm/hydrologic year), and a decreasing trendwas found atChittorgarh (−1.27 mm/hydrologic year) (Table S2). The per-centage change of significant trend was found to be at 102.85%and −80.29% in Hanumangarh and Chittorgarh, respectively(Table S2).

Significant decreasing trendswereobserved in annual averagerainfall for Jodhpur (at −7.51 mm/year and −85.05%), Karauli(at −7.41 mm/year and −34.65%), Nagaur (at −6.47 mm/yearand −61.93%) and Tonk (at −10 mm/year and −54.54%). In

Fig. 2. Location of urban centers with increasing, decreasing and no trends at 10% significance level for a) monsoon, b) non-monsoon and c) annual average rainfall.

80 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

81S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

the case of annual average rainfall, maximum values ofdecreasing slopes (−10 mm/year) and percentage change(−85.05%) were observed in Tonk and Jodhpur, respectively(Table S3). The IPCC also reported increases in precipitation at aglobal level, however, in the present study, precipitation wasfound to increase in some urban centers and decrease in othercenters. This may be due to changing land use/land cover, andseveral other anthropogenic activities at a local scale (Gowda etal., 2008).

4.1.2. Minimum temperatureSignificant increasing trends were observed in minimum

temperature at seasonal and annual temporal scales for themajority of the urban centers in the northeastern part ofRajasthan State. However, significant trends were not found inother urban centers at a 10% level of significance by the MK test(Fig. 3). The magnitude of the Sen's slope of significant trends inseasonal (monsoon, non-monsoon and annual) minimumtemperature for the urban centers of Rajasthan State is shownin Fig. 3. The maximum value of Sen's slope (0.04 °C/hydrologicyear) and percentage change (5.07%) for a significant increasingtrend of average minimum temperature in the monsoon seasonwere observed in Churu and Sikar, respectively. In the case ofnon-monsoon averageminimum temperature,maximumvaluesof increasing (0.05 °C/hydrologic year) slope were observedin Bikaner, Jaipur and Sikar. Similarly, percentage change innon-monsoon minimum temperature varied from 3.91% atBharatpur to 13.45% at Sikar. Also, a significant trend in annualaverage minimum temperature had maximum values of in-creasing slopes (0.05 °C/year) at Sikar. The percentage changesin annual average minimum temperature were found to be at aminimum (2.61%) and maximum (9.52%) at Banaswara andSikar, respectively. The positive trends of seasonal and annualminimumtemperatureswere also observed by other researchersin India (e.g., Jain et al., 2013).

4.1.3. Average temperatureSignificant increasing trends in average temperature at

seasonal and annual temporal scales were observed for themajority of the urban centers in Rajasthan State, accounting foran average of 80% of the urban centers (Fig. 4). The significantincreasing trends in monsoon average temperature wereobserved in Bikaner, Bundi, Churu, Dausa, Jaipur, Jaisalmer,Kota and Sikar, which is indicated by the maximum value ofslope (i.e., 0.03 °C/hydrologic year). Also, percentage change inmonsoon average temperature was found to be at a maximumfor Jaisalmer (3.47%) and at a minimum for Alwar (2.43%).Similarly, significant trends in non-monsoon and annual averagetemperature showed themaximumvalue of 0.03 °C in the urbancenters. The percentage changes in significant trends innon-monsoon average temperature were observed to be at aminimum (2.43%) and at a maximum (4.84%) in Banaswara andSikar, respectively. The percentage change in annual averagetemperature was found to be at a minimum (1.53%) andmaximum (4.32%) in Dungarpur and Sikar, respectively. Thedetail magnitude of the Sen's slope of the significant trends inaverage temperature is shown in Fig. 4. This warming trend inaverage temperature is also supported by other studies in India(e.g., Singh et al., 2008a, b).

4.1.4. Maximum temperatureSignificant increasing trends in monsoon average maximum

temperature were found in Barmer, Bikaner, Dausa, Jaisalmer,Jalore and Pali (Fig. 5a). The significant increasing trends inmonsoon season averagemaximum temperaturewere observedwith a maximum value (0.03 °C/hydrologic year) of Sen's slopein Jaisalmer and Bikaner. Percentage change in average maxi-mum temperature during the monsoon period was at aminimum (2.09%) and maximum (2.90%) in Jalore and Bikaner,respectively. In the case of non-monsoon average maximumtemperature, the southwestern urban centers in Rajasthan Statehad significant increasing trends. No significant trend was foundin the remaining urban centers during the non-monsoon periodat a 10% level of significance (Fig. 5b). Significant increasingtrends were observed in annual average maximum temperaturefor urban centers that are located in the southeast andsouthwestern parts of Rajasthan State (Fig. 5c). The significantincreasing trends in non-monsoon season average maximumtemperature were observed with a maximum value of Sen'sslopes (0.03 °C/hydrologic year) in Barmer, Jaisalmer, Jalore, Pali,Sirohi and Udaipur. Percentage changes in non-monsoonaveragemaximum temperature were found to be at aminimum(2.08%) and maximum (3.31%) in Dungarpur and Pali, respec-tively. No significant trendswere found in the averagemaximumtemperature at an annual temporal scale for the urban centerslocated in the northern region of the state (Fig. 5c). Similarly,significant trends in annual average maximum temperaturewere observedwithmaximum values of slopes (0.03 °C/year) inBarmer, Jaisalmer, Jalore, Pali and Udaipur. The percentagechange was found to be at a minimum (1.47%) and at amaximum (3.24%) in Pali and Baran, respectively. The detailmagnitude of the Sen's slope for averagemaximum temperaturefor the urban centers of Rajasthan State is shown in Fig. 5.These results are in good agreement with previous studies inIndia (e.g., Singh et al., 2008a, b).

4.2. Trends in extreme events

4.2.1. Extreme annual daily rainfallThe spatial distribution of extreme annual maximum daily

rainfall trends and variation is shown in Fig. 6a,while theMK teststatistics (Zmk) with increasing, decreasing and no trends areshown in Fig. 6b. Both positive (southeastern) and negative(northeastern) trends were observed in extreme annual dailyrainfall in the urban centers of Rajasthan State. The resultsrevealed that significant decreasing trends and variability wereobserved in annual extreme daily rainfall for Hanumangarh(at−0.65 mm/year and−49.16%), Karauli (at−1.56 mm/yearand −60.31%), Nagaur (at −1.56 mm/year and −80.33%),Sikar (at −1.33 mm/year and −63.86%) and Tonk (at−1.64 mm/year and −65.14%). However, a significant increas-ing trend was found only in Jhalawar (at 1.38 mm/year and45.97%). The maximum values of decreasing (−1.64 mm/year)slopes and percentage change (−80.33%) in significant trends inannual maximum daily rainfall were observed in Tonk andNagaur, respectively (Table S4). Significant trends were notfound for extreme events of annual daily rainfall in other urbancenters (Fig. 6). The detail magnitudes of slope and percentagechange in extreme annual maximum daily rainfall are presentedin Table S4 and Fig. 6. Similarly, only significant trends in extremeannual maximum daily rainfall are shown in Fig. 7. Investigating

Fig. 3. Location of urban centers with increasing, decreasing and no trends for a) monsoon, b) non-monsoon and c) annual average minimum temperature.

82 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

Fig. 4. Location of urban centers with increasing, decreasing and no trends at 10% significance level for a) monsoon, b) non-monsoon and c) annual average temperature.

83S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

Fig. 5. Location of urban centers with increasing, decreasing and no trends for a) monsoon, b) non-monsoon and c) annual average maximum temperature.

84 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

Fig. 6. Location of urban centers with increasing, decreasing and no trends at 10% significance level for extreme annual daily rainfall.

85S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

trends in extreme events is necessary to assess the impact ofglobal climate change over regional climatic variability onpresent and future water resources (Wang et al., 2013). Theassessment of trends of extreme annual daily rainfall events iscrucial for adaptation planning in arid and semi-arid regions ofIndia such as Rajasthan.

4.2.2. Extreme annual daily minimum temperatureThe spatial distribution of extreme annual daily minimum

temperature is shown in Fig. 8a, while the MK test statistic(Zmk) indicating trends is shown in Fig. 8b. Significantincreasing trends were observed in extreme annual dailyminimum temperature (at 0.03 to 0.07 °C/year) for the urbancenters located in the northeastern and western regions ofRajasthan State (Fig. 8). However, significant trends were notfound in other urban centers of Rajasthan State (Fig. S1)

(Supporting material). A significant increasing trend inannual minimum daily temperature was observed with amaximum value of slope (0.073 °C/year) and percentagechange (102.76%) in Sikar, while significant increasingannual daily minimum temperature was observed with aminimum value of slope (0.040 °C/year) and percentagechange (31.01%) in Banaswara and Hanumangarh, respec-tively. The details of slope magnitude and variability ofextreme annual daily minimum temperature in the urbancenters of Rajasthan State are presented in Table S5 andFig. 8. Only the trends in significant extreme events of annualdaily minimum temperature are shown in Fig. S1. From theresults, it is clear that the magnitude of the minimumtemperature trends and variability is greater than themaximum temperature extremes. This was also observed inother parts of India (Duhan et al., 2013).

y = -0.5473x + 1134.5 R² = 0.09070

102030405060708090

100

1970 1977 1984 1991 1998 2005

Ann

ual m

ax. d

aily

rai

nfal

l, m

m

Year1970 1977 1984 1991 1998 2005

Year1970 1977 1984 1991 1998 2005

Year

1970 1977 1984 1991 1998 2005

Year1970 1977 1984 1991 1998 2005

Year1970 1977 1984 1991 1998 2005

Year

HanumangarhRainfallLinear (Rainfall)

y = 1.7368x -3347.5 R² = 0.1206

0

50

100

150

200

250

300

350

Ann

ual m

ax. d

aily

rai

nfal

l, m

m

JhalawarRainfallLinear (Rainfall)

y = -2.3587x + 4779.6 R² = 0.1561

0

50

100

150

200

250

300

350

400

Ann

ual m

ax. d

aily

rai

nfal

l, m

m

KarauliRainfallLinear (Rainfall)

y = -1.2439x + 2540.6 R² = 0.0493

0

50

100

150

200

250

300

350

Ann

ual m

ax. d

aily

rai

nfal

l, m

m

NagaurRainfallLinear (Rainfall)

y = -1.298x + 2653.4 R² = 0.1839

0

20

40

60

80

100

120

140

160A

nnua

l max

. dai

ly r

ainf

all,

mm

SikarRainfallLinear (Rainfall)

y = -2.2845x + 4629.9 R² = 0.3097

0

50

100

150

200

250

Ann

ual m

ax. d

aily

rai

nfal

l, m

m

TonkRainfallLinear (Rainfall)

Fig. 7. Trends in statistically significant extreme annual daily rainfall in the urban centers of districts of Rajasthan State.

86 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

4.2.3. Extreme annual daily maximum temperatureThe spatial distribution of extreme annual daily maximum

temperature (obtained using MK test statistic (Zmk)) wasdetermined using the IDW interpolation technique for the 33urban centers of Rajasthan State (Fig. 9). A significant increasingtrend in extreme annual daily maximum temperature wasobserved in the south (i.e., Banaswara, Baran, Bundi, Chittorgarh,Jhalawar, Pali and Sirohi), east (i.e., Churu and Jaipur) andwest (i.e., Jaisalmer) urban centers of Rajasthan State (Fig. 9).Significant increasing trend in the annual maximum dailytemperature was observed with a maximum value of Sen'sslope (0.049 °C/year) and percentage change (3.86%) inJhalawar, while the minimum value of slope (0.029 °C/year)and percentage change (2.35%) was observed in Sirohi. Theminimum and maximum annual daily maximum tempera-tures were found to be at 44.48 °C and 47.62 °C in Sirohi andDholpur, respectively. The detail magnitude of the Sen'sslope and percentage change of annual daily maximumtemperature in urban centers of Rajasthan State is presentedin Table S6. The results of significant extreme events ofannual daily maximum temperature are shown in Fig. S2.The results clearly indicate the warming trends in theextreme annual maximum daily temperature in the urbancenters, which shows attribution of anthropogenic influenceon climate (Fig. 9a and b). Thesemay include drastic changesin land use/land cover, industrialization and urbanization.

In summary, themean and extreme rainfall and temperaturetrends and variability at a spatial and temporal scale wereevaluated for the 33 urban centers of Rajasthan State. Theresults of mean and extreme annual daily rainfall showedboth positive and negative trends over the urban centers. An

increasing trend was found in annual and seasonal mean andextreme minimum, and average and maximum temperaturesover the urban centers. This may be due to several natural andanthropogenic activities such as land use/land cover, urbaniza-tion and industrialization, which contribute to the changes inrainfall and temperature in terms of their distribution, intensityand occurrence in space and time. The detailed evaluation of themagnitude and variability in rainfall, and minimum andmaximum temperatures in terms of mean and extremes, isvery important for impact assessment and adaptation planningfor floods, drought and extreme events. It is suggested that inthe future, the assessment of mean and extreme eventsconsidering climate change scenarios can be carried out onseasonal and annual temporal scales at each urban center ofRajasthan State, India. This would be very useful to facilitate theproper planning and management of water resources in achanging climate in Rajasthan State, India.

5. Conclusions

In the present study, the trends in mean (monsoon season,non-monsoon season and annual) and extreme annual dailyrainfall and temperature were determined for the 33 urbancenters of Rajasthan State. This was carried out using thenon-parametric Mann–Kendall (MK) test. The Sen's slope andpercentage changes in rainfall and temperature were alsoestimated over the study period (1971–2005). Further, thespatial variations in mean (seasonal and annual) and extremeannual daily rainfall and temperature with their trends weredetermined using the inverse-distance-weighted (IDW) inter-polation method.

Fig. 8. Location of urban centers with increasing, decreasing and no trends at 10% significance level for extreme annual daily minimum temperature.

87S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

Both positive and negative trends were observed by the MKtest and Sen's Slope estimator in seasonal and annual averagerainfall in the urban centers of Rajasthan State. All of thesignificant trends in monsoon and annual rainfall were found tobe decreasing at the 10% level of significance. However, signif-icant trends in non-monsoon average rainfall were observed(both positive and negative trends). The significant decreasingtrends of average monsoon rainfall had the highest value ofdecreasing slope (i.e., −8.56 mm/hydrologic year) for Tonkand percentage change of −88.16% in Jodhpur. A significantincreasing trend was observed in non-monsoon rainfall atHanumangarh (1.70 mm/hydrologic year), and a decreasingtrend was found at Chittorgarh (−1.27 mm/hydrologic year).In the case of annual average rainfall, maximum valuesof decreasing slopes (−10 mm/year) and percentage change(−85.05%) were observed in Tonk and Jodhpur, respectively.

Significant increasing trends were observed in minimumtemperature at seasonal (0.04 and 0.05 °C/hydrologic year) andannual (0.05 °C/year) temporal scales for the majority of theurban centers in the northeastern part of Rajasthan State.Significant increasing trends (at 0.01 to 0.05 °C/year) wereobserved in average temperature for most of the urban centers(about 80% of the urban centers), in particular, all the urbancenters showed significant warming trends in annual averagetemperature. Significant increasing trends in monsoon averagemaximum temperature were found in the some of the urbancenters. However, the southwestern urban centers had signifi-cant increasing trends in non-monsoon average maximumtemperature. Significant increasing trends were observed inannual average maximum temperature for urban centers thatare located in the southeast and southwestern parts of RajasthanState. The trends in maximum and minimum temperature may

Fig. 9. Location of urban centers with increasing, decreasing and no trends at 10% significance level for extreme annual daily maximum temperature.

88 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

be influenced in different ways due to local physical geographic(e.g., topography, urbanization, air pollution etc.) and atmo-spheric circulation features (Turkes and Sumer, 2004; Dhorde etal., 2009; Tabari and Talaee, 2011a).

In the case of extreme events for annual daily rainfall bothpositive (southeastern) andnegative (northeastern) trendswereobserved in some of the urban centers. Other urban centers didnot show any significant trends in extreme annual daily rainfall.Significant increasing trends were observed in extreme annualdaily minimum temperature (at 0.03 to 0.07 °C/year) for theurban centers located in the northeastern andwestern regions ofRajasthan State. However, significant trends were not foundin the other urban centers of Rajasthan State. A significantincreasing trend in extreme annual dailymaximum temperature(at 0.03 to 0.05 °C/year) was observed in the south (i.e.,Banaswara, Baran, Bundi, Chittorgarh, Jhalawar, Pali and Sirohi),

east (i.e., Churu and Jaipur) and west (i.e., Jaisalmer) urbancenters of Rajasthan State. The results of mean (monsoon,non-monsoon and annual) and extreme annual daily rainfall andtemperature clearly indicate prominent changes in the urbancenters, whether due to natural variability or anthropogenicactivity over the period. The trends and variability of rainfall andtemperature found in this study could be one part of a cyclewitha frequency shorter than 35 years. A similar observationwas alsomade for precipitation trends in Iran (Tabari and Talaee, 2011a).There is no doubt that changes in natural factors like sea surfacetemperature, the Earth's rotation and solar cycles can causechanges in the climate. However, we do not have any control ofthese factors. Anthropogenic activities like urbanization and landuse change can also affect local climates significantly. The effectsof these factors could be minimized by taking appropriateadaptation and mitigation measures. The assessment of mean

89S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

and extreme events of rainfall and temperature is necessary tohelp and prepare suitable adaptation strategies in the face ofuncertain changing climate and extreme weather events. Theoutcome of this study should prove to be useful to policymakers,hydrologists, and water resource planners dealing with climatechange in all the urban centers in Rajasthan State for thesustainable development and planning of water resources.

Acknowledgments

We acknowledge the Indian Meteorological Department(IMD) in Pune for providing useful meteorological griddeddatasets for this researchwork. The financial assistance providedby the Ministry of Human Resource Development, Governmentof India in the form of a scholarship is duly acknowledged.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.atmosres.2013.10.024.

References

Arora, M., Goel, N.K., Singh, P., 2005. Evaluation of temperature trends overIndia. Hydrol. Sci. J. 50 (1), 81–93.

Basistha, A., Goel, N.K., Arya, D.S., Gangwar, S.K., 2007. Spatial pattern oftrends in Indian sub-divisional rainfall. Jalvigyan Sameeksha 22, 47–57.

Basistha, A., Arya, D.S., Goel, N.K., 2009. Analysis of historical changes inrainfall in the Indian Himalayas. Int. J. Climatol. 29, 555–572.

Cheung, W.H., Senay, G.B., Singh, A., 2008. Trends and spatial distribution ofannual and seasonal rainfall in Ethiopia. Int. J. Climatol. 28, 1723–1734.

Coulibaly, P., Shi, X., 2005. Identification of the Effect of Climate Change onFuture Design Standards of Drainage Infrastructure in Ontario, FinalReport. McMaster University, Department of Civil Engineering, Ontario.

Dash, S.K., Saraswat, Vaishali, Panda, S.K., Sharma, N., 2013. A study of changes inrainfall and temperature patterns at four cities and correspondingmeteorological subdivisions over coastal regions of India. Glob. Planet.Chang. http://dx.doi.org/10.1016/j.gloplacha.2013.06.004.

De, U.S., Rao, G.S.P., 2004. Urban climate trends — the Indian scenario.J. Indian Geophys. Union 8 (3), 199–203.

Deka, R.L., Mahanta, C., Pathak, H., Nath, K.K., Das, S., 2012. Trends andfluctuations of rainfall regime in the Brahmaputra and Barak basins ofAssam, India. Theor. Appl. Climatol. http://dx.doi.org/10.1007/s00704-012-0820-x.

Dhorde, A., Dhorde, A., Gadgil, A.S., 2009. Long-term temperature trends atfour largest cities of India during the twentieth century. J. IndianGeophys. Union 13 (2), 85–97.

Duhan, D., Pandey, A., 2013. Statistical analysis of long term spatial andtemporal trends of precipitation during 1901–2002 at Madhya Pradesh,India. Atmos. Res. 122, 136–149.

Duhan, D., Pandey, A., Gahalaut, K.P.S., Pandey, R.P., 2013. Spatial andtemporal variability in maximum, minimum and mean air temperaturesat Madhya Pradesh in central India. C. R. Geosci. 345, 3–21.

Erdoğan, S., 2009. A comparison of interpolation methods for producingdigital elevation models at the field scale. Earth Surf. Proc. Land. 34,366–376.

Gadgil, A., Dhorde, A., 2005. Temperature trends in twentieth century atPune. India Atmos. Environ. 39, 6550–6556.

Ghosh, S., Luniya, V., Gupta, A., 2009. Trend analysis of Indian summermonsoon rainfall at different spatial scales. Atmos. Sci. Lett. 10, 285–290.

Gosain, A.K., Rao, S., Basuray, D., 2006. Climate change impact assessment onhydrology of Indian river basins. Curr. Sci. 90, 346–353.

Gowda, K.K., Majuantha, K., Manjunath, B.M., Putty, Y.R., 2008. Study ofclimate changes at Davangere region by using climatological data. WaterEnergy Int. 65 (3), 66–77.

Hamlet, A.F., Mote, P.W., Clark, M.P., Lettenmaier, D.P., 2005. Effects oftemperature and precipitation variability on snowpack trends in theWestern United States. J. Clim. 18, 4545–4561.

Helsel, D.R., Hirsch, R.M., 2002. Statistical methods in water resourcestechniques of water resources investigations. Book 4, Chapter A3. U.S.Geological Survey. (522 pages).

Huang, J., Sun, S., Zhang, J., 2013. Detection of trends in precipitation during1960–2008 in Jiangxi Province, Southeast China. Theor. Appl. Climatol.http://dx.doi.org/10.1007/s00704-013-0831-2.

IPCC, 2001. Third assessment report. The Scientific Basis, vol. 1. CambridgeUniv. Press.

IPCC, 2007. AR4 working group I report. The physical science basis.Isaaks, E.H., Srivastava, R.M., 1989. An Introduction to Applied Geostatistics.

Oxford University Press, New York.Jain, S.K., Kumar, V., Saharia, M., 2013. Analysis of rainfall and temperature

trends in Northeast India. Int. J. Climatol. 33, 968–978.Jeganathan, A., Andimuthu, R., 2013. Temperature trends of Chennai City,

India. Theor. Appl. Climatol. 111, 417–425.Jiang, F., Hu, R., Wang, S., Zhang, Y., Tong, L., 2013. Trends of precipitation

extremes during 1960–2008 in Xinjiang, the Northwest China. Theor.Appl. Climatol. 111, 133–148.

Kampata, J.M., Parida, B.P., Moalafhi, D.B., 2008. Trend analysis of rainfall inthe headstreams of the Zambezi River Basin in Zambia. Phys. Chem.Earth 33, 621–625.

Karpouzos, D.K., Kavalieratou, S., Babajimopoulos, C., 2010. Trend analysis ofprecipitation data in Pieria region (Greece). Eur. Water 30, 31–40.

Kendall, M.G., 1975. Rank Correlation Methods, 4th ed. Charles Griffin,London.

Kumar, V., Jain, S.K., Singh, Y., 2010. Analysis of long-term rainfall trends inIndia. Hydrol. Sci. J. 55 (4), 484–496.

Ludwig, W., Serrat, P., Cesmat, L., Esteves, J.G., 2004. Evaluating the impact ofthe recent temperature increase on the hydrology of the Tˆet River(Southern France). J. Hydrol. 289, 204–221.

Luo, Y., Lio, S., Sheng, L., Fu, S., Liu, J., Wang, G., Zhou, G., 2008. Trends ofprecipitation in Beijing River Basin, Guangdong Province, China. Hydrol.Process. 22, 2377–2386.

Mall, R.K., Gupta, A., Singh, R., Singh, R.S., Rathore, L.S., 2006. Water resourcesand climate change: an Indian perspective. Curr. Sci. 90 (12), 1610–1626.

Mann, H.B., 1945. Non-parametric test against trend. Econometrica 13,245–259.

McBean, E., Motiee, H., 2008. Assessment of impacts of climate change onwater resources: a long term analysis of the Great Lakes of NorthAmerica. Hydrol. Earth Syst. Sci. 12, 239–255.

Modarres, R., da Silva, R.V.P., 2007. Rainfall trends in arid and semi-aridregions of Iran. J. Arid Environ. 70, 344–355.

Motiee, H., McBean, E., 2009. An assessment of long-term trends inhydrologic components and implications of water levels in LakeSuperior. Hydrol. Res. J. 40 (6), 564–579.

Nalley, D., Adamowski, J., Khalil, B., 2012. Using discrete wavelet transformsto analyze trends in streamflow and precipitation in Quebec and Ontario(1954–2008). J. Hydrol. 475, 204–228.

Nemec, J., Gruber, C., Chimani, B., Auer, I., 2012. Trends in extremetemperature indices in Austria based on a new homogenized dataset.Int. J. Climatol. 33, 1538–1550.

Pal, I., Al-Tabbaa, A., 2010. Long-term changes and variability of monthlyextreme temperatures in India. Theor. Appl. Climatol. 100, 45–56.

Pal, I., Al-Tabbaa, A., 2011. Assessing seasonal precipitation trends in Indiausing parametric and non-parametric statistical techniques. Theor. Appl.Climatol. 103, 1–11.

Pant, G.B., Kumar, K.R., 1997. Climates of South Asia. John Wiley, UK 320.Patra, J.P., Mishra, A., Singh, R., Raghuwanshi, N.S., 2012. Detecting rainfall

trends in twentieth century (1871–2006) over Orissa State, India. Clim.Chang. 111 (3), 801–817.

Rai, R.K., Upadhyay, A., Ojha, C.S.P., 2010. Temporal variability of climaticparameters of Yamuna River Basin: spatial analysis of persistence, trendand periodicity. Open Hydrol. J. 4, 184–210.

Rajeevan, M., Bhate, J., 2008. A high resolution daily gridded rainfall data set(1971–2005) for mesoscale meteorological studies. National Climate Centre,India Meteorological Department, Pune, Research Report No: 8/2008.

Rathor, M.S., 2005. State level analysis of drought policies and impacts inRajasthan, India. Colombo, Sri Lanka: IWMI. 40p. Working paper 93:Drought Series Paper 6.

Robinson, T.P., Metternicht, G., 2006. Testing the performance of spatialinterpolation techniques for mapping soil properties. Comput. Electron.Agric. 50, 97–108.

Roy, S.S., Balling Jr., R.C., 2004. Trend in extreme daily precipitation indices inIndia. Int. J. Climatol. 24, 457–466.

Saboohi, R., Soltani, S., Khodagholi, M., 2012. Trend analysis of temperatureparameters in Iran. Theor. Appl. Climatol. 109, 529–547.

Sahoo, D., Smith, P.K., 2009. Hydroclimatic trend detection in a rapidlyurbanizing semi-arid and coastal river basin. J. Hydrol. 367, 217–227.

Sen, P.K., 1968. Estimates of the regression coefficient based on Kendall's tau.J. Am. Stat. Assoc. 39, 1379–1389.

Singh, P., Kumar, V., Thomas, T., Arora, M., 2008a. Basin-wide assessment oftemperature trends in Northwest and Central India. Hydrol. Sci. J. 53 (2),421–433.

90 S.M. Pingale et al. / Atmospheric Research 138 (2014) 73–90

Singh, P., Kumar, V., Thomas, T., Arora, M., 2008b. Changes in rainfall andrelative humidity in river basins in Northwest and Central India. Hydrol.Process. 22, 2982–2992.

Some'e, B.S., Ezani, A., Tabari, H., 2012. Spatiotemporal trends and changepoint of precipitation in Iran. Atmos. Res. 113, 1–12.

Srivastava, A.K., Rajeevan, M., Kshirsagar, S.R., 2008. Development of a highresolution daily gridded temperature data set (1969–2005) for theIndian region. National Climate Centre, India Meteorological Depart-ment, Pune, Research Report No: 8/2008.

Tabari, H., Talaee, P.H., 2011a. Analysis of trends in temperature data in aridand semi-arid regions of Iran. Glob. Planet. Chang. 79, 1–10.

Tabari, H., Talaee, P.H., 2011b. Temporal variability of precipitation over Iran:1966–2005. J. Hydrol. 396, 313–320.

Tabari, H., Somee, B.S., Zadeh, R.M., 2011. Testing for long-term trends inclimatic variables in Iran. Atmos. Res. 100, 132–140.

Theil, H., 1950. A rank-invariant method of linear and polynomial regressionanalysis. I, II, III. Nederl. Akad. Wetensch. Proc. 53, 386–392.

Tozer, C.R., Kiem, A.S., Verdon-Kidd, D.C., 2012. On the uncertaintiesassociated with using gridded rainfall data as a proxy for observed.Hydrol. Earth Syst. Sci. 16, 1481–1499.

Turkes, M., Sumer, U.M., 2004. Spatial and temporal patterns of trends andvariability in diurnal temperature ranges of Turkey. Theor. Appl.Climatol. 77, 195–227.

Wang, H., Chen, Y., Xun, S., Lai, D., Fan, Y., Li, Z., 2013. Changes in dailyclimate extremes in the arid area of Northwestern China. Theor. Appl.Climatol. 112, 15–28.

Yavuz, H., Erdogan, S., 2012. Spatial analysis of monthly and annualprecipitation trends in Turkey. Water Resour. Manag. 26 (3),609–621.

Yevjevich, V., 1971. Stochastic Processes in Hydrology. Water ResourcesPublications, Fort Collins: CO.

Yue, S., Hashino, M., 2003. Long term trends of annual and monthlyprecipitation in Japan. J. Am. Water Resour. Assoc. 39 (3), 587–596.

Yue, S., Pilon, P., 2004. A comparison of the power of the t test, Mann–Kendall and bootstrap tests for trend detection. Hydrol. Sci. J. 49 (1),21–37.

Zhang, X., Vincent, L.A., Hogg, W.D., Niitsoo, A., 2000. Temperature andprecipitation trends in Canada during the 20th century. Atmos. Ocean 38(3), 395–429.

Zhang, X., Hogg, W.D., Mekis, F., 2001. Spatial and temporal characteristics ofheavy precipitation events over Canada. J. Clim. 14, 1923–1936.

Zhang, X.B., Zwiers, F.W., Li, G.L., 2004. Monte Carlo experiments on thedirection of trends in extreme values. J. Clim. 17, 1945–1952.

Zin, W.Z.W., Jamaludin, S., Deni, S.M., Jemain, A.A., 2010. Recent changes inextreme rainfall events in Peninsular Malaysia: 1971–2005. Theor. Appl.Climatol. 99, 303–314.