Inter-annual and spatial variation of altitudinal increase in rainfall over Mount Inthanon and Mae Chaem Watershed, Northern Thailand

: To understand the inter-annual and spatial variation of altitudinal increase in rainfall amount, duration and intensity over Doi (Mount) Inthanon and Mae Chaem Watershed, Northern Thailand, we have conducted rainfall observations in this area since 1997. Ten years of continuous data for 18 rain gauge stations were accu-mulated. Altitudinal increases in rainfall were observed for all nine years. Annual rainfall hours also had an altitudinal increase, whereas mean rainfall intensity had no clear altitudinal increase for the nine year period. Not only altitudinal rainfall increase but also inter- annual rainfall fluctuation was determined from the variations of rainfall hours. There was no difference in the intensity of altitudinal increases in rainfall between the western and eastern slopes of Doi Inthanon.


INTRODUCTION
Mountainous watersheds in the Asian monsoon region have important roles for people living downstream in terms of water resources and natural disaster prevention. Rainfall in mountainous regions is a crucial factor determining the possibility of water resource shortages, flooding, landslides and debris flow. The operational rainfall observation sites were distributed in an easily accessible downstream area; however, a rainfall observation site in a mountainous watershed area was lacking.
We established a rain gauge network in a mountainous watershed in Mae Chaem Watershed, Northern Thailand since September 1997 to understand the spatial and temporal variability of rainfall. Some preliminary results from these observations have been published by Dairaku et al. (2000) and Kuraji et al. (2001). Dairaku et al. (2000) reported the altitudinal increase in rainfall in this watershed using data from 1 June until 23 November, 1998 obtained by 13 rain gauges. Kuraji et al. (2001) also examines rainfall in this watershed using data from June 1998 until October 1999. They found that the altitudinal increase in rainfall was obvious in the two wet seasons in 1998 and 1999 with the increment in 1999 being more than 2.5 times greater than in 1998. The altitudinal increase in rainfall was also obvious in the dry season from 1998 to 1999, but the increment was smaller than those in the two wet seasons. It was not the rainfall intensity but the rainfall observed hours which causes the difference of altitudinal increment. Dairaku et al. (2004) reported the contribution of rainfall frequency, intensity and duration of several rainfall categories to the rainfall amount in the watershed using data from June to October for 1998 and 1999. However, these reports were based upon less than 2 years data from just 13 rain gauges and the universality and robustness of those findings have not yet been tested.
The objective of this paper is to present new findings that were not shown in previous papers by using 10 years of continuous data collected from 18 rain gauge sites. We sought to present (1) inter-annual fluctuations of rainfall and altitudinal increases in rainfall, and (2) a comparison of altitudinal rainfall increases between east and west slopes of the mountain range. T: Ota Keiki Co., 20 cm orifice diameter and 0.5 mm per tip) with pulse count time-recording data loggers (KADEC-PLS II: KONA System Co., one second time resolution). We installed two sites in September and seven sites in November 1997; four sites in May 1998; and two sites in November 1999. All 15 sites were located in the Mae Chaem watershed. To compare the east and west sides of the mountain range, two more rain gauges were installed on the east slope of Doi Inthanon in May 2004 (No. 18 and 19) and another rain gauge in November 2005 (No. 17). The locations, names and altitudes of these 18 rain gauge sites are shown in Figure 1 and Table I (Kuraji et al., 2007). The No. 16 site is a river discharge monitoring site and is therefore not shown in Table I. This watershed has been intensively studied in terms of small watershed hydrological modeling studies (Lim et al., 2004, Thanapakpawin et al., 2007, Vongtanaboon et al., 2008 Figure 2 shows the relationship between annual rainfall and altitude from 1999 to 2007. Altitudinal increases in rainfall are clearly shown for all years. The correlation coefficient between annual rainfall and altitude was higher than 0.87 for all years. Annual rain hours also had altitudinal increases (Figure 2). Average rainfall intensity, however, had no altitudinal increase and was almost constant (2 2.5 mm/hour) for all years. These relationships have already been reported by Kuraji et al. (2001) using data taken over less than 2 years; the present study confirmed these relationships were universal for 9 years. At site No. 14, average rainfall intensity in 2000 was 3.59 mm/hour, relatively high compared with the values of the other sites and in other years. We investigated and found that high intensity rain events were observed three times in September 2000 at site No. 14. These were 12.5 mm of rain for 8.35 minutes on the 11th, 13.5 mm for 13.1 minutes on the 21st and 34.0 mm for 8.35 minutes on the 24th of September. Except for these events, heavy rainfall events rarely occurred in this watershed during the observation period and at other sites.

Inter-annual variation of altitudinal increase in rainfall
Inter-annual fluctuations of rainfall at each site Figure 2 also shows that annual rainfall, annual rain hours and average rainfall intensity had inter-annual fluctuations. Figure 3 shows inter-annual variations in annual rainfall and annual rain hours at 3 sites (No. 3, 7, and 8) from 1999 to 2007. From Figure 3, it can be seen that rainfall and rain hour had a similar fluctuation pattern. The correlation coefficients between annual rainfall and annual rain hours were greater than 0.59 for all 15 sites in the Mae Chaem watershed. Only one site had a smaller correlation coefficient between interannual fluctuation in annual rainfall and that of annual rain hours than that of average rainfall intensity. Not only altitudinal rainfall increase, but also inter-annual fluctuation in rainfall was determined by annual rainfall hours rather than average rainfall intensity.
To understand the influence of rainfall on the interannual fluctuation in rainfall, inter-annual variations in annual rainfall and weak rainfall (rainfall intensity less than 5 mm/hour) at 3 sites (No. 3, 7, and 8) from 1998 to 2007 were shown in Figure 4. Dairaku et al. (2004) suggest that weak rainfall intensity events contribute to the altitudinal rainfall, and we found here that the interannual fluctuation is also influenced by the inter-annual fluctuations of the frequency of weak rainfall intensity events.   Figure 2. Altitudinal increase in annual rainfall (left), annual rainfall hours (center) and average rainfall intensity (right). average rainfall intensity (right) between the west (site Nos. 6, 8, 10, 12) and the east (site Nos. 17,18,19) sides of the mountain range. Because the rain gauges on the east side of the mountain were installed in May 2004, only data from 2005 to 2007 were used. Rainfall at Doi Inthanon (No. 7) is also shown for reference. From Figure 5 we found that there was no clear difference in altitudinal rainfall increase in the annual rainfall and annual rainfall hours between the west and east sides of the Doi Inthanon mountain range. In monsoon regions dominant wind reverses with seasons, and consequently upwind and downwind sides also reverse. As for Doi Inthanon, the eastern slope is the downwind side in summer. Figure 6 is the same as Figure 5 but only for May-June-July (MJJ) rainfall when the predominant wind direction is southwest. From Figure 6 we found that there was no clear decrease in rainfall along the eastern (downwind) side and conclude that no rain shadow effect appeared around Doi Inthanon during the summer rainy season.

Spatial variations of altitudinal increase in rainfall
The east side of Doi Inthanon has suffered from a serious water shortage (Walker, 2003) and people believed the area is in the so-called rain shadow of Doi Inthanon. This study shows that there are no differences between the east and west sides of the mountain in terms of rainfall; suggesting there might be another factor affecting the water resources shortage problem, which occurred mainly on the eastern side of the mountain .

CONCLUSION
This study confirmed the universality of previous studies on altitudinal increases in rainfall, rain hours and average rainfall intensity in the Mae Chaem water-20    (Nos. 17,18,19) are located on the east slope of Doi Inthanon, whereas the other sites (Nos. 6,7,8,10,12)  shed. Ten years of data accumulation enabled us to present the inter-annual fluctuations of rainfall, rain time, average rainfall intensity and altitudinal increases. Although recent advances in rainfall observation through remote sensing systems such as radar or satellite are remarkable, difficulties still remain, especially for mountainous areas with complex topography. A ground based rain gauge network is still essential to study rainfall characteristics in mountainous watersheds.
Natural water-related disasters such as flooding, landslides and debris flows, as well as social conflicts over water resources between upper and lower dwelling inhabitants of a watershed are steadily becoming more serious in South East Asia. One reason the watershed management system of a government might not work effectively to resolve such conflicts is the paucity of hydro-meteorological data for evaluation of heavy rains, water supply, consumption and shortages, due to an insufficient hydro-meteorological observation network. This study shows an example of a dense rain gauge network in a mountainous area where almost no hydrometeorological information had been available before. Such observation networks will be essential for watershed management to realize disaster mitigation and to resolve social conflicts over water resources in the South East Asian region.
This study focuses on inter-annual variation in altitudinal increase in rainfall but the seasonal variation of altitudinal increase in rainfall is the other important topic that needs to be studied. Okumura et al. (2003) examined the diurnal variation in rainfall observed by meteorological radar at Om Koi in Northern Thailand and found the rainfall characteristics were influenced by wind direction, which reverses with monsoonal transition. Synoptic disturbances such as tropical storms may also play a role in the inter-annual fluctuation in rainfall in this region. To carry out such studies we must prepare a combination of observational data (rain gauge network, meteorological radar and satellite) and reanalysis data. The region in this study is one of the possible sites where a combined data set will be available. We expect that much more hydro-meteorological research will be conducted over this site in the near future leading to applications for solving local societal issues.