Estimation of precipitation condensation latent heat in rainy season over Qinghai-Tibet Plateau

Transcription

1 Online system, Sciences in Cold and Arid Regions 2009, 1(2): Estimation of precipitation condensation latent heat in rainy season over Qinghai-Tibet Plateau DongLiang Li 1,2*, Hui Wang 1, Miao Liu 3 1. NIM, Nanjing University of Information Science and Technology, Nanjing, Jiangsu , China. 2. Cold and Arid Regions Environmental and Engineering Research Institute, CAS, Lanzhou, Gansu , China. 3. Climate Center of Zhejiang Province, Hangzhou, Zhejiang , China. *Correspondence to: DongLiang Li, NIM, Nanjing University of Information Science and Technology, Nanjing, Jiangsu , China. lidl@nuist.edu.cn Received: 20 December 2008 Accepted: 25 January 2009 ABSTRACT The basic data for this research comprise the outgoing long-wave radiation (OLR) data observed by the United States National Oceanic and Atmospheric Administration (NOAA) series satellites from June 1974 through December 2005 over the area of 75º 105ºE and 25º 40ºN (totaling 91 grid zones when the horizontal resolution is 2.5º longitude by 2.5º latitude) and the monthly rainfall data recorded, from 1961 through 2005, by 93 conventional meteorological stations on the Qinghai-Tibet Plateau. Based on the research of the relation between rainfall and OLR and climate regionalization, a mathematic model was established for each region and grid zone, which is applied to estimate the monthly rainfall and then to estimate the monthly latent heat resulting from the condensation of precipitation year by year from 1961 through The results indicated that the multi-year average precipitation is mm and the condensation latent heat is J in the eastern part of the Qinghai-Tibet Plateau; the increasing rate of condensation latent heat is J/10a in the recent 45 years; that is to say, it will increase 1.2 percent in each decade. Furthermore, the total condensation latent heat and its variation rate in the Qinghai-Tibet Plateau are slightly larger than in the east to the plateau. Keywords: Qinghai-Tibet Plateau; rainy season; rainfall; OLR; condensation latent heat 1. Introduction The thermal effect of the Qinghai-Tibet Plateau (from now on called plateau) plays a unique role in the westerly jet s northward jump, the occurrence and development of vortex over plateau, and the influence on the weather systems of the eastern part of China. Once vortex over plateau shifts eastward out and merges with the Southwest China vortex, it affects the weather and climate and causes drought and flood damages in eastern China. In the flood season of 1954, the entire Yangtze River basin experienced an extraordinary flood that hadn t happened in a hundred years; this was the result of 13 times rainstorms successively in this period, and each rainstorm was affected by the vortex system (Hu et al., 1997). In 1998, the middle and lower reaches of Yangtze River were flooded once again; eight times the river reached flood peak, and all were rare in history. The rainstorms in upper reaches of Yangtze River, which caused the third through eighth flood peaks, were closely related to the vortex system in the eastern part of the plateau (Yu, 2000; Deng et al., 2001; Li, 2002). The thermal effect of the plateau on East Asia atmospheric circulation during the course of seasonal conversion was researched (Zheng et al., 1999); this research showed that the plateau s thermal effect accelerates the southern branch jet stream s northward jump obviously in early summer, but the pure dynamic action actually weakens it remarkably, even causes its southward retreat after northward jump. Obviously, the plateau s thermal effect has major impact on the East Asia atmospheric circulation conversion process in early summer. Moreover, the influence of the plateau s dynamic and thermal effects on global and tropical atmospheric circulation during the course of seasonal conversion was nu-

2 100 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): merically studied (Zheng et al., 1999 and 2001); that research indicated those items play a vital role in simulating the global and tropical atmospheric circulation correctly. The ground-effective radiation and latent and sensible heat are the main means of heat-transport from surface to atmosphere. The plateau s thermal forcing to the atmosphere gives priority to the sensible heat in the dry season. With the precipitation increase, however, the latent and sensible heat may reach the same magnitude in rainy season; at that point, the latent heat becomes an important energy of the plateau to affect the atmospheric circulation (Ye et al., 1979; Zhang et al., 1988). We need to gather data about the rainfall of the survey region regardless of the research of the plateau climate change or ecological environment evolution. Unfortunately, gathering it is not practical. Because the plateau is scarcely populated, an insufficient number of observation stations have been established to gain the precipitation data. The satellite material, however, has the characteristic of global observation, which might make up for the shortage of observation stations on the plateau. In the early 1970s, Barrett (1970) proposed that using satellite cloud picture to estimate the precipitation of the sea surface; this was the beginning of estimating rainfall with satellite material. After that, with the improvement of satellite data, many researches in this field took place one after another. Some research was took place using the data of outgoing long-wave radiation (OLR), which are observed by means of satellite remote sensing; the value of that data, however, largely depends on the temperature of the irradiator (surface, cloud top). OLR is mainly influenced by the cloud under homogeneous situations in large-scale low- and middle-latitude areas, and it has a close relation with clouds, and may also respond with precipitation information. Therefore, the use of OLR to estimate precipitation of a tropical ocean had already been successful (Arkin, 1983) and has since been put into use in estimating rainfall (Climate Diagnostics Bulletin, 1984). A great deal of research was also done in China using OLR. For example, an experiment was carried out to estimate the rainfall in middle and lower reaches of the Yangtze River (Jiang et al., 1986); a series of 10-year OLR data was used to estimate the monthly rainfall and condensation latent heat of the Qinghai-Tibet Plateau in summer (Xu et al., 1990). Other research and analyses included the relation of the climate characteristics of OLR. with precipitation of regions in Northeast China during summer (An et al., 1998); the OLR distribution difference in summer of the drought and flood years over the Qinghai-Tibet Plateau (Jia et al., 2002), and climate characteristics of the plateau OLR and their influences on Northern Hemisphere atmospheric circulation (Li et al., 1996). Generally speaking, although the relations between OLR and precipitation have some minor differences in different regions, there is a good anti-correlation between them. However, the existing research has some disadvantages; for instance, the results are many years old, the data used is short, the precipitation observation stations are few on the plateau, and so on. Thus, this could not clearly and definitely express the climatic features. This article uses monthly OLR and rainfall data of the plateau in the most recent 30 years. The relation between OLR and precipitation was further analyzed on the base of Liu s work (2007), and the total rainfall and latent heat sequences has been gained through the dependent equation of the rainfall and OLR in the most recent 45 years. 2. Data and technique This article uses the OLR data from June 1974 through December 2005 (but with a lack of data for March through December, 1978). This data was observed by U.S. National Oceanic and Atmospheric Administration (NOAA) series satellites. Those satellites altogether provide 91 grids in the scope of 75º 105ºE and 25º 40ºN when the horizontal resolution is 2.5º longitude by 2.5º latitude. The source of these data is ftp://ftp.cdc.noaa.gov/ Datasets/interp_OLR/. Also used was data from 1961 through 2005 providing the monthly rainfall of 93 conventional meteorological stations on the plateau. Figure 1 shows the locations of the OLR point grids and precipitation observation stations. Figure 1 The locations of the OLR point grids ( + ) and precipitation observation stations ( ) on the plateau. (Liu et al., 2007) Precipitation sequence R(t), t=1,2,, n, can use a linear regression equation R(t) =a OLR(t)+b, to estimate with OLR data. According to the least-square method, one may obtain coefficient a and constant b. 3. Relationship between rainfall and OLR and climate regionalization in the rainy season on the plateau Figure 2 shows the distribution of the 30-year average OLR (solid line, interval 50 W/m 2 ) and precipitation (dashed line, interval 50 mm) in the rainy season on the plateau. In that figure, one can see that the average OLR value is low in the east and high in the west in the rainy

3 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): season of the plateau, and its low value is less than 205 W/m 2, and the center is in 95º 100ºE along 30ºN. According to Liu s analysis, the OLR low-value center has obvious seasonal variation characteristics: when the sky is clear in the partly cloudy dry season, the value of OLR is mainly decided by the surface temperature, and the surface temperature is determined by altitude and snow. Therefore, the OLR low-value center is in Karakoram Mountains in the dry season (December and January). After the winter, along with the gradual evolution from spring to summer, convective activity strengthens, and cloud amount increases gradually under the influence of thermal-forcing on the plateau; plus, the OLR low-value center shifts from west to east, finally arriving at the plateau central place in May. The precipitation in the east is more than that in the west, which is opposite with OLR. This is primarily controlled by the warm-wet air stream of the Bay of Bengal; the precipitation center is also in 95º 100ºE along 30ºN. The low- (high-) value center of precipitation is essentially coincident with OLR, and between them remains an obvious anti-correlation in the rainy season of the plateau. Figure 2 OLR (solid line, interval 50 W/m 2 ) and precipitation (dashed line, interval 50 mm) over the plateau from May through September From this, we may obtain the regression equation through the interrelation of OLR and precipitation; then according to the OLR data estimates can be made for the rainfall and condensation latent heat on the western part where no plateau observation stations are located. To estimate the precipitation and precipitation condensation latent heat of the plateau, we primarily used the OLR data from 1974 through 2005, which had 46 grid points in the scope of 77.5º 102.5ºE and 27.5º 37.5ºN. We also employed the 1961 through 2005 rainfall data of 93 observation stations on the plateau. The precipitation and divisions of OLR were considered in order to better estimate the precipitation condensation latent heat of the plateau. The plateau s main body is divided into three districts (depicted by the heavy line): DistrictⅠis the humid region, Ⅱis the semi-arid region, and Ⅲ an arid area. According to the geographic position and precipitation distribution, the plateau is divided into eastern and western parts (expressed by a dashed line). Figure 3 shows that each grid zone has observation station on eastern part of the plateau, and most of stations were established before 1961; moreover observation stations are scarce on western part of the plateau, and the majority of grid zones have no station. 4. Estimation of precipitation condensation latent heat in the rainy season on the eastern part of the plateau There, on eastern part of the plateau, are even distributed observation stations, at least each grid zone has a station, and most of the stations were established before So according to Xu s appraisal results (1990), as long as we have one or more observation station s precipitation data, its representation is sufficient for 2.5º 2.5º longitude and latitude regions. No observation station region s rainfall can be revised through the method of OLR estimation. Therefore, the condensation latent heat of the eastern plateau is mainly based on the rainfall of each grid calculated; the western part is estimated through the relation of precipitation and OLR, because the observation stations are few and observation time is short. The precipitation condensation latent heat s calculating formula is: Latent heat = mean rainfall of the grids (cm) district s area (cm 2 ) ρ L V In this formula, ρ is the specific gravity of water that is 1 g/cm 3, and L V is the condensation latent heat coefficient, namely, L V =2,497 J/g, and its unit is joule (J).

4 102 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): Estimation of average precipitation in grid zone In the course of estimating condensation latent heat, the computation of the average precipitation has two kinds of situations on grid zones. Studies have shown that the representation of two observation station s precipitation data is sufficient for 2.5º 2.5º longitude and latitude regions. Along with an increase in the number of stations, the correlation of precipitation and OLR will be better, but the correlation coefficient increases indistinctly, and approaches to a constant gradually. Moreover, the correlation coefficient was not much different when there was only one station (Xu, 1990). In summary, if the grid zone has observation stations, the mean rainfall is the arithmetic mean of all precipitation data; if the grid zone has no observation station, according to the climate regionalization principle [that is to say, the precipitation (OLR) of various stations (grid zone) have the similar variation characteristics in the same climatic region], the mean rainfall can be estimated with the regression equation between the average precipitation and OLR departure of this climatic region through using the OLR departure value to estimate the precipitation departure value to eliminate the systematic climate variability. Figure 3 Distribution of the grids and observation stations on the plateau. The thick line is for climate regionalization, and the dashed line is for eastern and western parts Computation of grid zone area The latitude and longitude lengths of each grid: Each longitude length is: cosφ (unit: km) Each latitude length is: (unit: km) Because OLR data s horizontal resolution is 2.5º longitude by 2.5º latitude, the area computation formula for various grid zones is: ( cosφ cosφ 2 2.5) / Thus, the various grid zone areas respectively are: 1 to 9 region: cm 2 10 to 20 region: cm 2 21 to 31 region: cm 2 32 to 40 region: cm Estimation of latent heat resulting from the condensation of precipitation The correlations of average precipitation and OLR are calculated, and then the rainfall estimation model is established on the grid zones that had no observation station. Finally, each month s rainfall and the precipitation condensation latent heat of all grid zones on the plateau are obtained. Table 1 shows the correlation coefficient (r) and regression coefficients (a, b) of the precipitation and OLR in three divisions. It can be seen that most correlation coefficients are at 0.01 confidence level, except in May in the Ⅲ-region (arid area) and from May through September of the Ⅱ-region (semi-arid region) are below 0.05 confidence level. This is mainly because the western part plateau has not yet entered the rainy season in May, and in September, the northeastern plateau s rainy season has ended, dry season has begun, and the precipitation events are few and unsteady.

5 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): Table 1 Correlation coefficient (r) and regression coefficients (a, b) between the OLR and precipitation on climate regions of the plateau from May through September Region May Sept. May June r a b r a b r a b Ⅰ Ⅱ Ⅲ Region July Aug. Sept. r a b r a b r a b Ⅰ Ⅱ Ⅲ Table 2 Average rainfall (AR) (mm) and condensation latent heat (CLH) ( J) over eastern part of the plateau from May through September Year May June July Aug. Sept. May Sept. AR CLH AR CLH AR CLH AR CLH AR CLH AR CLH Average

6 104 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): Because the amount of precipitation has a close relation with OLR, which would guarantee the regression equation s reliability between them, and guarantee that precipitation and condensation latent heat can be estimated accurately for entire plateau, especially in the rainy season on the eastern part of the plateau. Thus, the monthly average rainfall and precipitation condensation latent heat for eastern part of the plateau from 1961 through 2005 (Table 2) and the entire plateau from 1975 through 2005 (Table 3) are estimated according to the amount of precipitation s observed value and the estimated value used OLR data. Table 3 The average rainfall (AR) (mm) and condensation latent heat (CLH) ( J) on the plateau from May through September Year May June July Aug. Sept. May Sept. AR CLH AR CLH AR CLH AR CLH AR CLH AR CLH Average Figure 4 shows the change curve of total latent heat on eastern part (a) and entire (b) of the plateau from May through September. The analysis indicates that annual average precipitation is mm, and the condensation latent heat is J on the eastern part of the plateau. It increased a bit in the most recent 45 years, and the increasing rate is J/10yr. That is to say, it had increased 1.2 percent in each decade. Furthermore, the annual variation of condensation latent heat is very obvious on eastern part of the plateau, and its highest (lowest) value is J ( J) in 1998 (1994), which is about J ( J) more (less) than the annual average, that is to say, it increases (decreases) by 14.3 percent (16.4) percent, the difference is J (30 percent) between them. Preceding text pointed out that precipitation condensation latent heat is on the increase in the most recent 45 years, but there has been an obvious decreasing process for more than 10 years from the 1960s to the early 1970s; after that, it increased slowly and steadily, which agrees with the increase of the amount of net water vapor on "Three River Sources Areas" in the most recent 40 years (Li, 2009). Because the precipitation observation is short, and the

7 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): observation stations are few, the precipitation condensation latent heat estimated is only from 1975 through 2005 on the western plateau 15 years less than on the eastern end. Table 2 shows that the plateau s overall annual average precipitation is only mm, and the condensation latent heat is J, which is equal to eastern plateau s maximum value. The increasing rate of precipitation condensation latent heat is J/10yr; that is to say, it will increase 2.83 percent in each decade, which is greater than in the eastern part of the plateau. This happens mainly because the average precipitation radix is small in northeastern portion of the plateau, so the increase appears a bit larger. Furthermore, the highest (lowest) value of condensation latent heat is J ( J) in 2000 (1994), which is about J ( J) more (less) than the annual average; that is to say, it increases (decreases) by percent (11.0 percent), the difference is J (32.1 percent) between them, which is a little bigger than that on the eastern part of the plateau (30 percent). This further explains the precipitation s instability in the western part of the plateau. Figure 4 The change curve of precipitation condensation latent heat ( J) in the eastern part (a) and the entirety (b) of the plateau from May through September 5. Conclusions and discussions The amount of precipitation has a good correlation with OLR observed by satellite in the rainy season on the plateau; through the relation of the two, one can estimate the total rainfall and precipitation condensation latent heat of the plateau. The annual mean precipitation increases a bit in the most recent 45 years, and the increasing rate is J/10yr; that is to say, it will increase 1.2 percent in each decade. Furthermore, the total precipitation condensation latent heat and its change rate of the plateau are a little larger than that on the eastern part of the plateau.

8 106 DongLiang Li et al., 2009 / Sciences in Cold and Arid Regions, 1(2): REFERENCES An G, Jiang SC, OLR climate characteristics and the relationship with the precipitation in summer on northeast China. Meteorology of Jilin Province, (2): Arkin PA, A diagnostic precipitation index from infrared satellite imagery. Tropical Ocean Atmos., Newsletter, 17: 5 7. Barrett EC, The estimation of monthly rainfall from satellite data. Mon. Wer. Rew., 101: Climate Diagnostics Bulletin, Climate analysis center, U.S. NOAA, NWS, NMC. Dong WJ, Wei ZG, Fan LJ, Climatic character analysis of snow disasters in east Qinghai-Xizang Plateau livestock farm. Plateau Meteorology, 20(4): Hu AG, Lu ZC, Sha WY, Chinese Natural Disaster and Economic Development. Science Press of Hubei Province, Wuhan, China. Jia L, Zhou SW, OLR distribution in summer in dryness/wetness years over Tibetan Plateau. Journal of Applied Meteorological Science, 13(3): Jiang SC, Winston JS, The characteristic of outing long wave radiation related to flood and drought over Yangtze River basin. WMO /TD, 87: Li DL, Zhang JJ, Wu HB, A study on the effects of climatic feature of over the Qinghai-Xizang Plateau on general circulation in northern hemisphere. Plateau Meteorology, 5(3): Li SC, Li DL, Zhao P, The water-vapor transfer characteristic in rainy season of the source area of "Three Rivers" over Qinghai-Xizang Plateau. Acta Meteorologica Sinica, 67(6): (in press). Li YQ, Analyses of cloudiness, sunshine, temperature and daily range on the eastern side of Qinghai-Xizang Plateau in recent 40 years. Plateau Meteorology, 21(3): Liu M, Li DL, Change characteristic and correlation of OLR and precipitation over east Qinghai-Xizang Plateau in rainy season. Plateau Meteorology, 26(2): Xu GC, Li DL, Jiang SC, Evaluating of monthly rainfall and condensation latent heat over Qinghai-Xizang Plateau in summer utilizing the OLR data of satellite. Plateau Meteorology, 9(3): Ye DZ, Gao YX, Meteorology of Qinghai-Xizang Plateau. Science Press, Beijing, China. Yu SH, An analysis of impact of the heavy rain in upper reaches of the Yangtze River on the flood peak of the river in Meteorology, 26(1): Zhang JJ, The Aadvances of Meteorology on Qinghai-Xizang Plateau. Science Press, Beijing, China. Zheng QL, Liang F, Numerical study of the effects of dynamic and thermal of Qinghai-Xizang Plateau on global atmospheric circulation during the course of seasonal conversion. Journal of Tropical Meteorology, 15(3): Zheng QL, Wang SS, Zhang CL, Numerical study of the effects of dynamic and thermodynamic of Qinghai-Xizang Plateau on tropical atmospheric circulation in summer. Plateau Meteorology, 20(1): Zheng QL, Wu J, A numerical study on the effect of the Qinghai-Xizang Plateau on the seasonal transition in the early summer in East Asia. Acta Meteor Sinca, 9(1):