Temperature Distribution and Air Pollution in Phnom Penh , Cambodia-Influence of Land Use and the Mekong and Tonle Sap Rivers

The temperature distribution in Phnom Penh was measured using a car to evaluate the thermal characteristics and air pollution in the city. The measurements were made by using temp-hygro sensors with a data logger installed on the car roof and were conducted both during day and night to evaluate the influence of land use. The water temperature was also measured in the River Tonle Sap and the Mekong. By measuring the temperature as a function of distance from the banks of the Mekong and Tonle Sap, the cooling effect of the river was also investigated. Ambient particulates were simultaneously sampled at three different sites in the city along with NO2 and PAHs and heavy metal concentrations were analyzed. The temperature distribution was compared with the concentration of chemical compositions and NO2. The maximum temperature difference, the so called heat island intensity, was observed during the daytime and was around 4-5C, and was less than 2C during the nighttime. The maximum and minimum temperatures respectively were observed in the southern part of the city and the river peninsula between the Tonle Sap river and the Mekong, and a strong cooling effect of river water was found. The water temperature was consistently lower than the ambient and temperature distributions perpendicular to the river and was found to increase with the distance from the riverbank, suggesting that inland areas were cooled to some extent. Comparison of the concentrations of anthropogenic PAHs and NO2 were found to be closely related to temperature. keywords: Urban air, temperature measurement, air pollutants, river temperature, GIS. * Corresponding author. Tel: +81-76-234-4646; Fax: +81-76-234-4646. E-mail address: mfuru@t.kanazawa-u.ac.jp Furuuchi et al., Aerosol and Air Quality Research, Vol. 6, No. 2, pp. 134-149, 2006


INTRODUCTION
During past 10years after the election of 1993, the Cambodian economy has grown rapidly (JICA, 2001;National Institute of Statistics of Cambodia, 2006) and the urban area of Phnom Penh, the capital of Cambodia, has been also expanding, and a significant increase in energy consumption and the number of automobiles has occurred.This inevitably gives rise to an increase in air pollution as well as an increase in temperature, or, thermal pollution, which leads to health problems and an uncomfortable living environment (Oke, 1987).Since sources of emitted air pollutants and heat are frequently the same, air quality and heat pollution are closely linked with each other (Jonsson et al., 2004;Roberts, 2004).Although the available data are limited, a few investigations and measurements to the air pollution (Matsui, et al., 2002;Kashima et al., 2001) have been reported and these studies indicate a serious environmental situation.
However, thermal pollution has not been evaluated in Phnom Penh.
In this study, the temperature distribution inside Phnom Penh was measured using a car to evaluate the thermal characteristics and air pollution in the city.Measurements using two sets of temp-hygro sensors with a data logger installed on the car roof were conducted both during the day and night to evaluate the influence of various heat sources.The water temperature in Tonle Sap and Mekong rivers was also measured.Temperature distributions were evaluated in relation to the distance from the riverside in order to investigate the cooling effect of the river.Ambient particulates were simultaneously sampled in three different sites in the city along with NO 2 and PAHs and heavy metal concentrations were also analyzed.Temperature distribution was compared with the concentrations of chemicals and NO 2 .

Structure and land use of Phnom Penh
Fig. 1 shows land use data evaluated from satellite data (ASTER, AST3A1, 20/02/2002) using commercial GIS software ( ENVI ver.4.2 ) based on the supervised classification method (Tanaka et al., 2003;Johnson et al, 2003).Each type of land use is designated by 6 different gray levels (25(forest), 75(blank space 2), 100(water area), 130(blank space1), 200(paddy field) and 255(urban area)).The urban area on the left bank of the Tonle Sap and Mekong rivers, designated in white, is widely spread along radial streets from the central market and some main rings perpendicular to the radial streets: the busiest commercial area around the central market, intermingled commercial and residential areas spread to the west and south, governmental and commercial areas between Monivoung street and the riverbank, a residential area in north west, which are all spreading over a rather flat area of land.There are no skyscrapers but 6-8 stories buildings dominate the central area particularly along the main roads and lower buildings cover a large area of the city.As can be seen from Fig. 1, the area of vegetation occupies a very small portion but there are large rivers, Mekong, Tonle Sap and Bassac on the east side of the city, as well as rather large water ponds in north and south.There is heavy traffic on the main streets in the city each morning and evening.

Measurement of Ambient Air Temperature
The measuring probe consisted of a temperature-hygro sensor with a data logger (SATO KEIRYOKI SK-L200TH II) attached inside a PVC tube (55mm I.D., 300mm length and 2mm thickness) covered by an aluminum coated urethane foam sheet (6mm thickness) (See Fig. 2(a)).
The probe has almost the same geometry as one used in a previously published paper (Wong and Yu, 2005).To avoid the influence of the direct impingement of airflow, the axis of the probe was directed perpendicular to direction the car was traveling.For the compensation, two sets of observation probes were installed on both sides of the roof of the car, as shown in Fig. 2(b).
The focal points for the observations were as follows: 1) influence of land use on temperature distribution, 2) difference between the day and night temperature distribution and influencing factors, and 3) cooling effect of the rivers.The routes and conditions for the observation are 1, for night time observation along with the influence of distance from the river on temperature and Route-3 the influence of the distance from the river on temperature.
After leaving the starting point, or, the Diamond Hotel located in the central downtown area, the observing car passed though checkpoints A, B, C…, which were determined in advance, following the traffic lights.The average speed and total travel distance for each route are also shown in Table 1.Temperature and humidity were recorded at preset time intervals (10 sec.-1 min.)while the time was recorded at each checkpoint.The temperature was also measured at the top of the Diamond Hotel, or an 8 story building, to monitor changes in temperature and humidity during the observations.

Measurement of River Water Temperature
The water temperatures in both Tonle Sap and Mekong rivers were measured at the sites shown in Fig. 1.After boating to each site, the water temperature distribution from the bottom to the surface of the river was measured using three different types of thermometers: "multi-parameter water quality profiler (Alec Electronics Co. Ltd., type AQ1183 with 10 m cable)", which measures the vertical profile of dissolved oxygen (DO), pH, chlorophyll concentrations, temperature, turbidly and electrical conductivity of waters at intervals of 5cm.The others were "Compact CT Meter (Alec Electronics, Japan)" and "Hobo Water Temp Pro (Onset Computer, USA)".Data sampled every 10 or 20 minutes were stored in loggers connected to each thermometer.The temperatures were measured from Jan. 2005 to Dec. 2005 and the 1 st of August was the closest day to the air temperature measurement.Tanaka et al (2003) reported that the water level in August and September in Phnom Penh during past several years was almost constant.
According to the Mekong Secretariat (1994), the flow rate of the River Mekong in August is ca.
10% less than that in September and the average air temperatures are the almost same so that the difference in water temperature during these months would be expected to be small.Therefore, the river water temperature in August is used in the following discussion.

Measurement of Air Pollutants
Total suspended particulates (TSP) and NO 2 were simultaneously sampled at three different analyzed by means of a spectrometer following a previously reported procedure (Yanagisawa and Nishimura, 1980).

Difference and temperature distribution along main streets for day versus night
Figs. 4(a)-(c) show the temperature distributions respectively along Route-1, 2 and 3. Since the temperature variation during the observation was less than 0.3 o C, the temperature data were  This is almost on level with the heat island intensity and population for cities in Japan (Ishida and Katsumoto, 1999).
Over Route-2, the temperature difference was not as large as in the daytime but gradually Route-3, although there was a difference due to the difference in weather conditions, a similar tendency to Route-1 was observed.Temperature variations through the straight part of Monivoung Street (F￫O in Route-1 and 2), which is one of the main axes of the city, running almost parallel to the river, is shown in Fig. 5, where the difference between the temperature and the average temperature T av over each route is plotted against the distance from point F.During the nighttime, the temperature increased toward the central downtown area then reached a peak around L-N, followed by a decrease to the urban periphery.During the daytime, point A, which is located in the busiest area, was classified as being in a cooler area and a similar increase was found from the central to the southern downtown area (A￫M).The urban area of Phnom Penh has a rather uniform land use but the north of the busiest area has a less wall developed urban area (See Fig. 1) so that the center of the "heat island" should be located in the south.There might also be an influence of river cooling around point A, which will be described below.The higher temperature around point F during the daytime was possibly due to the heavy traffic through a traffic circle.
The temperatures at the check points along the river, particularly in the river peninsula (H-J) were consistently lower than the others.This may be due to the cooling effect of the rivers.In ).Thus, there might also be some river cooling effects in the nighttime.However, the temperature variation may be influenced to a greater extent by land use, since the temperature decreased in D/D M > 1 and the difference between T av and T r was small.indicating an increase in the pollutant concentration with the air temperature.These results may be an evidence for the temperature dependency of air pollutants concentration although it will need to be verified based on further observations.(1)The maximum temperature was observed, not in the busiest area, but in the southern part of the city both during the day and night.The maximum temperature difference, the so called the heat island intensity, was 4.6 o C, which is almost on a correlation with the intensity and population for cities in Japan.
(2)A clear influence of the Mekong and Tonle Sap rivers was found regarding the cooling of ambient temperature in the downtown area during the daytime, extending to around 2 km from the riverbank.
(3)The air temperature seemed to be more sensitive to land use during the nighttime while it may be more influenced by traffic and river cooling effects during the daytime.
(4)The wind over the rivers may dilute air pollutants and prevent their transportation from downtown to the river peninsula.
(5)The Mekong and Tonle Sap rivers have important roles in the thermal and environmental characteristics of the ambient atmosphere in Phnom Penh.

Fig. 1 .
Fig. 1.Structure and land use in Phnom Penh as evaluated from Satellite (ASTER) data, where air pollutant sampling sites (residential, central downtown and river peninsula) and the locations of water temperature measurements (P-T(Tonle Sap) and P-M(Mekong)) are also shown.

Fig. 2 .
Fig. 2. (a) Observation probe (a temp-hygro sensor with data logger), (b) observation probes installed on the roof top of a car.
sites in Phnom Penh respectively using high volume air samplers (SHIBATA HV-500F) and passive samplers (ADVANTEC Personal Passive Sampler: Filter badge NO 2 ).The sampling sites are shown in Fig. 1:1) Central downtown (the Diamond Hotel), 2) Residential area, 3) River peninsula between the River Macon and Tonle Sap.Quartz fibrous filters (ADVANTEC QR-100) were weighed after storing them in a constant humidity and temperature for 72 hours using a descicator.The sampling conditions are summarized in Table 2. Since the observations of temperature and air pollutants were done in different months but in the same rainy season, pollutant data can be discussed with temperature data.The filters were weighed after a 72 hour conditioning period then analyzed to determine the concentrations of heavy metals and polycyclic aromatic hydro-carbons (PAHs).Heavy metals were analyzed by ICP-AES after decomposing the samples with nitric, hydrofluoric and perchloric acids.15 different PAH compositions, that is, Naphthalene (Nap), Acenaphthene (Ace), Phenanthrene (Phe), Anthracene (Ant), Fluorene (Fle), Fluoranthene (Flu), Pyrene(Pyr), Benz[a]anthracene (BaA), Chrysene (Chr), Benzo[a]pyrene (BaP), Benzo[b] fluoranthene (BbF), Benzo[k]fluoranthene (BkF), Dibenz[a,h]anthracene(DbA), Indeno[1,2,3-cd]pyrene (IDP) and Benzo[ghi]perylene (BghiPe) were analyzed using HPLC with a fluorescence detector + acetonitril/ultra pure water mobile phase after ultrasonically dissolving the samples on the filter in an ethanol/benzene (1:3) solution and evaporation with a rotary vacuum evaporator.NO 2 was Figs. 4(b)  and (c), the average temperature of the Tonle Sap and Mekong rivers is also shown.Regarding Route-1, the temperature increased along the main road with two lanes (A￫B￫C) and decreased around the residential area (D) with more vegetation.The temperature then reached a peak in the traffic circle (F).The temperature drastically decreased down to the minimum over the Route-1 at the cape of the peninsula (I).Along the main road (Monivoung Street) running north to south, the temperature gradually increased through the central downtown (A) to the south (A￫M￫N￫L￫O￫P) and reached a peak (P￫Q).After these points, the temperature increase was not so clear in central downtown while it decreased near the river bank (V￫W￫X).A rapid decrease just before arriving at the starting point corresponded to the initiation of precipitation.The maximum temperature difference, so called "heat island intensity" was 4.6 o C.

Fig. 4 .
Fig. 4. Temperature distributions along (a) Route-1 (day time), (b) Route 2 (night time) and (c) Route-3 (day time) order to determine the area encompassed by the cooling effect, the temperature difference between air and river water, T -T r , is plotted against the longitudinal distance D from the riverbank in Figs.6(a)-(b), where T and T r are, respectively, air and river water temperatures and T -T r is normalized by the difference between average temperature T av over each observing route and T r ; the distance D is also normalized by the average distance between the straight section of Monivoung street (F-O) and the riverbank D M (~1.47km).During the daytime (Fig. 6(a)), the air temperature can be expressed in almost the same relation regardless of the observation day, where T abruptly increases near the riverbank from T r then gradually increases with distance.The cooled area may extend to D/D M ~ 1, which covers most of the area between Monivoung Street and the riverbank.At locations that are influenced more by traffic, denoted by open circles, the air temperature became slightly higher.Some locations influenced by traffic had lower temperatures, probably due to cooling by the river (Traffic&River).During the nighttime (Fig. 6(b)), T increased gradually from the riverbank and a clear peak in the temperature was noted around Monivoung Street (D/D M ~ 1

Fig. 5 .
Fig. 5. Variation in air temperature T along the main road (Monivoung Street) running from north to south through the Phnom Penh urban area, where T av is the average temperature for each route.

Fig. 6 .
Fig. 6.Difference between air and river water temperatures plotted against the distance from the riverbank of River Mekong, where the temperature difference is normalized by the difference between the averaged temperature T av along each route and the river water temperature T r .and the distance from the riverbank D is normalized by the average distance D M between Monivoung street along check points F to O.

Fig. 7 .
Fig. 7. Variation in normalized air temperature from the average temperature over each observation route plotted against the urban area fraction evaluated from ASTER data: (a) daytime and (b) nighttime

Fig. 8 .
Fig. 8. Concentrations of total suspended particulates (TSP) and total PAH compositions with four or more aromatic rings plotted against the day-averaged air temperature obtained at the site in the central downtown (Diamond Hotel, May 2005).