Urbanization Effect on Air Temperature in Iran's Zayanderud River Basin

Department of Irrigation and Drainage Engineering, College of Abureyhan, University of Tehran, Tehran, Iran Department of Natural Resources Engineering, Isfahan University of Technology, Isfahan, Iran Department of Water Engineering, Isfahan University of Technology, Isfahan, 8415683111, Iran Department of Civil Engineering, Isfahan (Khorasgan) Branch, Islamic Azad University, Isfahan, Iran Civil Engineering Department, South Tehran Branch, Islamic Azad University, Tehran, Iran


Introduction
Urbanization is a human activity that modifies the nature of the surface and atmospheric properties of a region. This phenomenon affects the surface energy budget, land-atmosphere interaction, hydrologic regimes, resulting in urban climate feature so-called Urban Heat Island (UHI) effect (Collier, 2006;Habete and Ferreira, 2016). More than half of the world's population is living in the cities and heat waves due to UHI can significantly affect human health. So, it is important to know the contribution of urbanization to the observed warming trends. Many studies were conducted on comparing rural and urban warming and most of them pointed out a warmer climate in urban areas due to UHI effect (Yan et al., 2010;Ren and Zhou, 2014). Jones et al. (2008) showed that temperatures in a suburb of London and Vienna increased relative to rural locations nearby in recent decades. Karl et al. (1988) studied the variation of temperature during 1901-1984 in urban and rural areas of United States and found that even small cities experienced the evident influence of urbanization on the temperature trend. Peterson et al. (1999) divided the 7280 stations of the world into rural, suburban and urban stations and, by analyzing the global mean temperature trend based on all of stations and the rural stations, they observed that the rural stations are warming at 0.8°C (100yr)-1 while the increase of temperature for all stations was 0.92°C (100yr)-1. Ren et al, (2008) by comparative analysis of rural and urban stations investigated urbanization effect in north China. They found that the impact of urban warming is 0.11°C (10yr)-1 and the contribution of urban warming to total temperature changes is 37.9%. Hamdi et al. (2009) analyzed the effect of urbanization in the Brussels capital region, Belgium.
They estimated annual mean urban bias and contribution of UHI effect to overall warming trend by 0.61°C and 45% respectively. Chrysanthou et al. (2014) studied the effect of urbanization in Europe since1960. The result of this study indicated 1.45% (0.0026°C/10yr) of overall temperature trend (0.179°C/10yr) might explain by urbanization effect. Ebrahim et al. (2016) investigated the effect of urbanization on the daily temperature in Dubai during 1980-2012. They showed that the average temperature region recorded has been steadily increasing with a high reading of 47°C Celsius in 2012. Almazroui et al. (2013) investigated the possible effects of urbanization on the rise of air temperature in Saudi Arabia for the period 1981-2010 using population data. They indicated warming trend by 0.5-0.6°C/decade but suggested that the rise in air temperature is not likely due to urbanization changes. Saadatabadi and Bidokhti (2011) analyzed temperature trend in Tehran megapolis and Varamin (a small town in the southeast of Tehran) as a rural station to investigate urbanization effect. The result of the study indicated that temperature trend varies in response to the different rate of urbanization. Abbasnia et al. (2016) studied the variation of temperature trend in Esfahan megacity, Iran during 1980-2010 and indicated that constructed area and temperature has increased 23% and 0.06°C respectively. Also, they found the significant correlation between heart moralities and temperature trend.
Several studies have done to detect temperature trend in Iran and Zayanderud basin (Tabari and Hosseinzadeh-Talaee, 2011;Gandomkar, 2011;Ataei et al., 2013;Ghasemi, 2015) but few researches (types of research) have been conducted on the effect of urbanization upon air temperature trend and it is unclear at present, how much effect the increased UHI magnitude has had on the long-term trends in Zayanderud basin.
Previous studies pointed out that increasing temperature in Zayanderud basin is an influential factor in imbalance between water consumptions and resources (Gohari et al., 2013;Zareian et al., 2015;Ahmadi et al., 2015) and it is expected that the findings of this study contribute to getting better insight into sources of regional warming and more efficient management in the basin.

Zayanderud River Basin
Zayanderud river basin with a semiarid climate is located in Isfahan province, central Iran (Fig. 1). The area of the basin is over 41,000 Km 2 with a population (of)over 5.3 million people where 70% of the population is living in the cities. The third megacity of Iran (Esfahan) located in the centre of the basin contains 50% of the basin's population.The Annual mean temperature of the basin is 13°C where east and west of the basin are the warmest and coldest areas (Fig. 2). Zayanderud dam in western basin stores runoff to supply urban, industrial, agricultural etc. demands.

Climatic Data
The monthly data during 1951-2010 were received from Iran meteorological organization and ministry of energy. There are 113 stations located in Zayanderud basin and around. Record lengths of these stations vary between 5 to 60 years. To enhance the validity of the results, the stations selected based on a large record of data (more than 30 years), continuity of the records and appropriate distribution throughout the basin (Ghahraman, 2006;Githui, 2009). On the basis of these criteria, 37 stations were selected for analysis in this study. All records started from the 1950s, 1960s, 1970s and 1980s to 2010. Among selected stations 21 stations, are located in the rural area and 16 stations are located in urban area. The location of stations is shown in Fig. 3. In this figure, urban and rural station is shown with "U" and "R", respectively. To avoid as much as possible, the influence of urbanization, rural stations were selected from villages, desert and mountaintop or towns with a population less than 5000 people (Ren et al., 2008;Ren and Zhou, 2014). Monthly values of stations were averaged to obtain annual temperature for each station. To analysis, the homogeneity of the records, double-mass curve (Kohler, 1949) was used for each station. The results of double-mass curve showed that there are no obvious breakpoints in time series. There were a few gaps in selected stations. To fill these gaps, the recorded data in neighbour stations with high correlation were used.

Methodology
In this study, in order to detect the magnitude of temperature trends, Q Sen's slope estimator was applied (Sen, 1968). This method is commonly used for hydrometeorological time series data. The slope estimates for N pairs of data are first calculated by: where, x j and x i are the data values at times j and i (j >i) respectively. The median of these N values of Q k is a Sen's estimator of slope: The pre-whitened time series was used prior to applying Q Sen's slope estimator for removing the influence of the serial correlation. In this study, at first serial correlation coefficient (r 1 ) is calculated for time series (x 1 , x 2 , x 3 ,…….,x n ) with lag-1 according to Equation 3: where, n is the number of observations and E (x i ) is the mean of sample data. If the calculated r 1 is significant, the Sen's estimator will be used on 'pre-whitened' time series and if r 1 is not significant original time series can be used. (Tabari and Hosseinzadeh-Talaee, 2011;Gocic and Trajkovic, 2013;Islam, 2015).
After computing the magnitude of trends in stations, spatial distributions of trends over the basin was applied in Geographical Information Systems (GIS) and the raster map of the trend by all (rural-urban) stations and by rural stations were generated. Generated map by rural stations would be used as the reference for preurbanization and the map generated by all stations was considered as temperature changes of post-urbanization. To compute UHI effect, the values of rural-urban raster layer were subtracted from the values of rural (or reference) raster layer. (Peterson et al., 1999;Ren et al., 2008;Ren and Zhou, 2014). In this study, the spatial variation of trends was determined using the inverse-Distance-Weighted (IDW) interpolation technique. The IDW is a simple and widely used method for interpolating hydro-meteorological variables. This method that is based on the assumption that the values of the variables at a point to be predicted are similar to the values of nearby observation points (Pingale et al., 2014). The formula of IDW (Bayazit et al., 2016) is as follows: where, Z * (x 0 ) shows predicted measure of x 0 point and Z (x i ) indicates the measure of sample point at x i point.
The w (d i ) is the weighting function and d i is the distance from x i to x 0 . p is the power factor and n is the number of sample points.

Trends in stations
The summarized outputs of Q Sen's Slop estimator for mean annual temperature trend in 37 stations is presented in Table 1. In this table increasing and decreasing trends are shown with positive and negative signs respectively. According to the results, 32 out of 37 stations have an increasing trend, 4 stations have experienced decreasing trend and 1 station is without trend. Tiran station in the centre of the basin (33-U) and Ghalesharokh station in the west of basin (14-R) have the highest and the lowest trends. Figure 4 shows the distribution of temperature trend in Zayanderud basin after urbanization. The average temperature trend over the basin is 0.012°C year-1 and the highest and the lowest magnitude of trends are seen in central and western basin respectively. According to Ghasemi (2015), the average warming trend in Iran is 0.023°C year-1 and in comparison to the country as a whole, the warming trend over Zayanderud basin is generally weaker. In addition to urbanization effect, the reason for regional differences in the magnitude of warming over the basin may refer to physical geographic (e.g., topography, exposure), vegetation, land cover, land use, atmospheric circulation etc. Further investigation is needed to find the relation between trends and these factors. Figure 5 shows the distribution of temperature trend in Zayanderud basin before urbanization. The average of temperature trend before urbanization was 0.010°C year-1. According to the findings, the weakest pre-urbanization trends are seen in the west of the basin and stronger trends occurred in a southeastern basin. Figure 6 presents the difference between pre and posturbanization warming (UHI effect) over the basin. The average UHI effect over the basin is 0.002°C year-1 so 17% of overall warming rate occurred in Zayanderud basin is caused by urban warming. According to information of ministry of the country, from 1970 to 2010, the number of cities in the basin increased from 31 to 77. In addition to changing the rural area to urban area, the already urbanized cities have expanded due to increasing population. Available information show in 1995, 2005 and 2010 urban population of the basin was 2.5, 3.2, 3.5 million people respectively. This urbanization process resulted in an increase of temperature over the basin. Sources of this warming refer to industrial processes, power plants, combustion process of vehicles, human metabolism and transfered heat from residential and commercial buildings. In addition, expansion of the cities from non-construction land to construction land with high buildings (absorb more radiation) leads to rise temperature (Sailor and Lu, 2004;Yan et al., 2016).

UHI Effect
The centre of the basin experiences strongest UHI effect by 0.006-0.012°C year-1. Figure 7 shows the distribution of the cities in Zayanderud basin and around. 45 out of 77 cites of the basin are located in the central basin. These cities contain 70% of the basin's population. In addition, the majority of urbanized cites of the basin including Esfahan, Khomeinishahr, Najafabad, Zarinshahr, Fuladshahr, Shahinshahr and Mobarake are located in the centre of basin. It appears centre of the basin due to more concentration of anthropogenic activities experiences larger UHI effect. In the east and southeast of the basin cities are seen less than other part of the basin and most population are living in rural areas with a low living standard. In addition, in comparison with other zones of the basin in these areas, the cities are less urbanized. These issues perhaps are the reason for the lower UHI effect in this area.  In the analysis of UHI effect, determination of rural stations as reference stations is a key issue. Choosing the rural stations which are less affected by urbanization led to use a limited number of rural stations in the interpolations. This may overestimate or underestimate UHI effect in some part of the basin. The 21 reference stations as used in this study was the best ones we were able to obtain at present. Development of the stations in the future might lead to more reliable outcomes.

Conclusion
This study aimed to evaluate urbanization effect on mean annual air temperatures in Iran's Zayanderud river basin using analysis of regional trends. The results showed that warming trends dominate over the basin and larger trends are observed in the central basin. Comparison of the pre and post-urbanization trends exhibited the increase of trends caused by urbanization effect. It was estimated that 17% of the overall warming trend in the basin is related to urbanization effect. The strongest effect of urbanization was seen in the centre of the basin where urbanized and the densely populated area is and the weakest urbanization was observed in the eastern and southeastern basin where the cities are less urbanized and the rural societies are dominated. The results presented in this study suggest that urbanization affect the overall warming in Zayanderud basin and may influence imbalance between water consumptions and resources. Further studies are needed to analysis the contribution of urbanization on water shortage in the basin.
In this study, we provide document for a considerable urbanization relic on climate depended on analysis of effects of land-use changes on surface temperature, where quick urbanization has happened. The importance of urban warming relics on zonal average temperature trends is evaluated applying monthly mean temperature series of the station group datasets, which tolerate in homogeneity modification. The spatial template and dignity of our survey are stable with those of urbanization determined by alterations in the percentage of urban population and in satellite-measured greenness. We analysis the surface air temperature record of the past 150 years, examining the concord of the primary data and the standard errors of approximation of the average hemispheric and global evaluates. The twentieth-century warming has been attended by a decline in those regions of the global influenced by exceptionally cool temperatures and to a lesser extent by enhances in areas influenced by exceptionally warm temperatures. A new collection of monthly average surface air temperature for the Northern Hemisphere for 1851-1984 is offered depend on land-based meteorological station data and fixed-position weather ship data. This new land-based hemispheric temperature curve is contrasted with last estimates of Northern Hemisphere temperatures depended on maritime data. In the universal lands, the bias of urbanization influences still exits in the surface air temperature series of many city weather stations to a certain extent. Credible reference climate stations require to be chosen for the finding and correction of the local manmade warming intolerance.

Conclusion
The main conclusions of the experimental work should be presented. The contribution of the work to the scientific community and its economic implications should be emphasized.