December 1992 M. Kawamata, S. Yamada, T. Kudoh, K. Takano and S. Kusunoki 1161 NOTES AND CORRESPONDENCE. Atmospheric Temperature Variation

Transcription

1 December 1992 M. Kawamata, S. Yamada, T. Kudoh, K. Takano and S. Kusunoki 1161 NOTES AND CORRESPONDENCE Atmospheric Temperature Variation after the 1991 Mt. Pinatubo Eruption By Masahiro Kawamata, Shingo Yamadal, Tatsuya Kudoh, Kiyoharu Takano Long-range Forecast Division, Japan Meteorological Agency, Tokyo 100, Japan and Shoji Kusunoki Numerical Prediction Division, Japan Meteorological Agency, Tokyo 100, Japan (Manuscript received 6 August 1992, in revised form 12 October 1992) 1. Introduction In June 1991 Mount Pinatubo (15*N, 120*E) erupted and injected volcanic ash and gases, including large amount of sulfur dioxide, into the stratosphere (Smithsonian, 1991). The sulfur dioxide was converted into a sulfuric acid aerosol in the stratosphere, which absorbed and scattered sunlight, and then affected atmospheric temperature. Ueno (1992) reported that the observed global solar radiation at Tsukuba in Japan showed a decrease of about 5% from normal in December It was also pronounced by Aerological Division of the Japan Meteorological Agency (JMA) based on rawinsonde data in Japan that the lower stratospheric temperature over Japan rose in the latter half of 1991 (JMA, 1992). From their global analyses of rawinsonde data, Labitzke et al. (1983), and Labitzke and McCormick (1992) reported stratospheric warming after the eruption of Mt. El Chichon in March 1982 and Mt. Pinatubo in June It was demonstrated by Hansen et al. (1978) from numerical simulations that the stratospheric temperature increased because of aerosol absorption of sunlight and upwelling long wave radiation. This note will report on atmospheric temperature variations after the eruption of Mt. Pinatubo using the daily data output from the operational global objective analysis of the Japan Meteorological Agency. 'Present affiliation: Forecast Division, Japan Meteorological Agency, Tokyo, 100, Japan. 1992, Meteorological c Society of Japan 2. Data The data used here are the daily (12UTC) temperature data from the JMA operational global analysis at 16 standard pressure levels from the surface to 10hPa, with latitude-longitude grids of 1.875* The data are available from January 1985 to August The thickness data retrieved from the report of Satellite Remote Upper-Air Soundings of Pressure, Temperature and Humidity (SATEM) are incorporated into the temperature analysis as observational data. Considering the possibility of a quality change of SATEM measurements due to the injected ash and gases of the volcanic origin, we have examined the reliability of SATEM data before and after the eruption of Mt. Pinatubo. For this purpose, the temperature estimated from the geopotential height of rawinsonde observation is used as a reference temperature. Data of SATEM are compared with those of the rawinsonde, only if the distance between the location of the two data is less than or equal to 150km and the observation time difference between the data is less than or equal to 3 hours. Figure 1.1 shows the vertical distribution of mean errors and root-mean-square errors of thickness temperature derived from SATEM thickness between two adjacent standard pressure levels for May Figure 1.2 shows a similar result for August Numbers of data pairs used in the calculation are entered for each level in the panel of MEAN errors. We can see that mean errors and root-mean-square errors are less than 1* and 3* respectively in all levels and latitudes. Comparing Fig. 1.2 with Fig. 1.1, it is found that the data qual-

2 1162 Journal of the Meteorological Society of Japan Vol. 70, No. 6 Fig The vertical distributions of mean errors and root-mean-square errors of thickness temperature derived from SATEM thickness between two adjacent standard pressure levels for May (a) errors from SATEM data located at the north of 30*N, (b) between 30*N and 30*S and (c) at the south of 30*S. Numbers of data pairs used in the calculation are listed for each level in the panel of mean errors. The unit are * (See Section 2 of text). ity of SATEM is essentially unchanged before and after the eruption of Mt. Pinatubo. Fig Same as Fig. 1.1 but for August The stratospheric temperature Figure 2 shows the analysed temporal change of global and zonal mean monthly temperature anomalies in the lower stratosphere. These anomalies are defined as deviations from the monthly mean temperatures averaged over the period from 1985 to The global mean temperature anomaly began to increase in July 1991 toward its maximum deviation *2* in October After that it decreased gradually and remained still positive in August The maximum anomaly in 1991 was larger than twice the maximum anomaly before 1990, although the available data were for only 8 years since Considering these large positive anomalies and the

3 *. December 1992 M. Kawamata, S. Yamada, T. Kudoh, K. Takano and S. Kusunoki 1163 Fig. 2. Time series of global and zonal mean monthly thickness temperature anomalies between 30hPa- 100hPa (thin line) and their 5-month running mean (thick line). Normal is the 6-year monthly mean. The uppermost shows global mean, the second shows 90*N-30*N zonal mean and so on. The units are fact that the anomalies began to increase just after the eruption of Mt. Pinatubo, we can conclude that the warming in the lower stratosphere in 1991 and 1992 was caused by the injection of volcanic aerosol from Mt. Pinatubo. The zonal mean stratospheric temperature anomaley in the latitude belt between 30*N and 30*S increased more sharply than those in the other belts (90*N-30*N, 30*S-90*S) after the eruption of Mt. Pinatubo. It decreased gradually after October The year-to-year variation of lower stratospheric temperature is also related to the phase of the Quasi-Biennial Oscillation (QBO). The equatorial wind flow near 30hPa changed from westerly to easterly in the Spring of If Mt. Pinatubo had not erupted in 1991, the temperature at 50hPa might have been below normal due to the phase of the QBO in the latter half of 1991 (Labitzke and McCormick, 1992). The increase of the mean temperature anomalies in the latter half of 1991 was not so apparent in the northern extra-tropical region as in the low latitudes. In the southern extra-tropical region, on the other hand, it was as large as in the low latitudes. It should be noted that the other volcano, Mt. Hudson (46*S, 73*W), erupted in August It might also affect the atmospheric temperature through aerosol injected into the stratosphere. However, the increase in temperature anomaly began before the eruptions of Mt. Pinatubo and Mt. Hudson. In addition, positive temperature anomalies as large as those in 1991 have occasionally appeared, i, e., in 1986 and They would be related with the stratospheric sudden warming in the southern hemisphere. Therefore we may not be able to conclude that the positive anomalies in 1991 in the southern hemisphere occurred only due to volcanic aerosol. Figure 3 shows the latitude-height cross section of monthly mean temperatures (broken curves) and their anomalies (solid curves). Areas of negative anomalies are hatched. The warming of the lower stratosphere became evident in August 1991, and positive anomalies greater than *2* spread between 30*N and 20*S in February The spatial distribution of 3-month (January-March 1992) mean temperature anomalies at 50hPa are shown in Fig. 4. Positive temperature anomalies dominated all over the latitudes south of 40*N. The warming was evident in the latitude belts around 20*N and 20*S, and was particularly evident around 20*N. 4. The Tropospheric Temperature It was also shown by Hansen et al. (1978, 1992) that an increase of volcanic aerosol in the stratosphere caused the tropospheric temperature fall through a decrease in short-wave radiation due to scattering and absorption. Figure 5 shows the analysed temporal change of tropospheric temperature anomalies similar to that of the stratosphere in Fig. 2. The fall of the global mean tropospheric temperature anomaly began in September 1991 and continued until January Then the global mean temperature anomaly increased slightly in spring and again showed a tendency to decrease in the Summer

4 1164 Journal of the Meteorological Society of Japan Vol. 70, No. 6 Fig. 3. Latitude-height cross sections of monthly mean temperatures (broken curves) and their anomalies (solid curves). Areas of negative anomalies are hatched. The units are *. (a) May 1991, (b) August 1991 (c) November 1991 (d) February 1992 (e) May 1992, (f) August Fig month (Jan.-Mar. 1992) mean temperature anomalies at 50hPa. Areas of negative anomalies are hatched. The contour interval is 1*. Normal is the 6-year ( ) monthly mean.

5 December 1992 M. Kawamata, S. Yamada, T. Kudoh, K. Takano and S. Kusunoki 1165 Fig. 5. Same as Fig. 2 but for 300hPa-850hPa. of But the magnitudes of the anomalies were as large as those of year-to-year variations before In all the latitude belts (90*N-30*N, 30*N- 30*S, 30*S-90*S), temperature anomalies decreased in the latter half of The 5-month running mean of the temperature anomalies of the 30*N- 30*S and 30*S-90*S belts turned to an increase in early 1992, but that of the 90*N-30*N belt continued to decrease. It is known that sea surface temperature affects the variation of tropospheric temperature. Angell (1988) showed that the global mean temperature anomaly increased with a lag when a warm episode of El Nino occurred. It must be noted that the global mean temperature decreased in spite of the warm episode of El Nino in 1991/92. Therefore it is likely that the decrease of temperature is due to volcanic aerosol in the stratosphere. Of course, the tropospheric temperature varies due to various other factors, and we must be careful to conclude these. 5. Conclusion We presented the atmospheric temperature variations after the eruption of Mt. Pinatubo as a prompt report. The lower stratospheric warming after the eruption of Mt. Pinatubo was evident. Tropospheric cooling was found, but the magnitudes of the anomalies were as large as those of year-to-year variations. For further discussion of this cooling, it is necessary to make a quantitative investigation of the radiative effect of aerosols. Acknowledgements The authors express their thanks to Director S. Yoshizumi and the colleagues of the Long-range Forecast Division for their encouragement. This work was a part of "Effects of the Pinatubo eruption on Climate" supported by the Science and Technology Agency. References Angell, J.K., 1988: Impact of El Nino on the delineation of tropospheric cooling due to volcanic eruption. J. Geophys. Res., 93, Hansen, J.E., W.C. Wang and A.A. Lacis, 1978: Mt. Agung eruption provides a test of global climatic perturbation. Science, 199, Hansen, J., A. Lacis, R. Ruedy and M. Sato, 1992: Potential climate impact of Mount Pinatubo eruption. Geophys. Res. Lett., 19, Japan Meteorological Agency, 1992: Monitoring of Global Warming and Ozone Depletion 1991 (In Japanese). Japan Meteor. Agency. 233 pp. Labitzke, K., B. Naujokat and M.P. McCormick, 1983: Temperature Effect on the Stratospheric of the April 4, 1982 eruption of El Chichon, Mexico. Geophys. Res. Lett., 10, Labitzke, K. and M.P. McCormick, 1992: Stratospheric Temperature Increases due to Pinatubo Aerosols. Geophys. Res. Lett., 19, Smithsonian Institution, 1991: Bulletin of the Global Volcanism Network Vol. 16, No. 5. Ueno, T., 1991: An analysis of the Solar Radiation Data at Tsukuba about the Eruption of Pinatubo Volcano in 1991 (In Japanese). Journal of the Aerological Observatory at Tateno, 52,

6 1166 Journal of the Meteorological Society of Japan Vol. 70, No. 6