SOUTHWEST CHINA: WHAT'S GOING ON WITH AIR QUALITY DATA?

UNCLAS SECTION 01 OF 02 CHENGDU 000095 SENSITIVE SIPDIS DEPT FOR EAP/CM E.O. 12958: N/A TAGS: SENV CH ENRG PREL ECON SUBJECT: SOUTHWEST CHINA: WHAT'S GOING ON WITH AIR QUALITY DATA? REF: 2008 CHENGDU 210 CHENGDU 00000095 001.2 OF 002 ¶1. (SBU) This cable contains sensitive but unclassified information - not for distribution on the Internet. ¶2. (SBU) SUMMARY: Air quality in Southwest China's city of Chengdu, is probably worse than official statistics would suggest. The city records significantly fewer poor air quality days than should normally occur based on a statistical analysis of the government's air quality data. Even air quality that has a strong odor and limits visibility is sometimes recorded as being only "lightly polluted." City residents in May reportedly called local media outlets to question the disparity between official air quality figures and air they saw outside. Defensive of their data, city environmental officials claim pollution is often localized and lasts for only short periods throughout the day. END SUMMARY. ¶3. (SBU) Air quality in Sichuan's capital city of Chengdu is on track to record more relatively low pollution days in 2009 than at any time since 2001, based on government data available for the first five months. Through May, Chengdu had 124 days with "very good" air quality, defined as days on which the air pollution index (API) is 100 or lower. Air quality monitors currently measure only sulfur dioxide, nitrogen dioxide, and large particulates, neglecting other pollutants like ozone and fine particulates. Chengdu may record 329 days of very good air by the end of the year, roughly 90 percent of time, if the pattern seen in our analysis of the data from previous years continues. Past data from the city's Environmental Protection Bureau shows that air quality generally improves after the winter months. Over the last five years, Chengdu's air quality statistics show that the number of very good quality air days increased 7 percent on average from June 1 through the end of the year. When "Choking" Air Equals Light Pollution -------------- ¶4. (SBU) Despite the apparent success and the potential for full-year improvements in air quality, the city's data appears to show the same potential for manipulation Post has seen in reviewing the last nine years of reported measurements. Days on which air quality probably did not reach the very good threshold were nevertheless recorded as "very good." We see an excessive number of days during the first five months of this year where pollution data falls outside the normal distribution one would expect to see. Through the end of May 2009, Chengdu recorded about 60 more days with an API of 46-100, and roughly 25 fewer days of API 100-150 than would occur if the data had a normal distribution. Note: Normal distribution, also called Gaussian distribution, has long been used as a basis for analyzing air quality. In a normal distribution, a large number of independent data points will tend to follow a bell curve specific to the data set being analyzed. Chengdu's air quality data has significant variations from this curve, particularly around the cutoff point between lightly polluted and very good air quality. ¶5. (SBU) The annual burning of agricultural stalks around Chengdu during May in fact resulted in noticeably poor air quality for the city's residents. The resulting pollution decreased visibility in the downtown area and gave the air a distinctive smell. A local Chinese-language news service, Sichuan Online, reported that on 18 May some city residents were calling the newspaper to ask about the difference between the published figures and the perceived air quality. The newspaper wrote that "pervasive smoke" on May 17 and 18 "choked people," but official figures show that the API for both days was 65, very good, and 104, lightly polluted, respectively. A Sichuan Evening News reporter said that on 17 May visibility in Chengdu's northwestern district of Jinniu was less than 100 meters, and drivers were forced to significantly reduce their speed while driving. A picture included with the news article showed a thick shroud of pollution in what appears to be downtown Chengdu. Congenoff also noted the smell of burning CHENGDU 00000095 002.2 OF 002 agricultural stalks south of the city center. The Chengdu city government officially prohibits the burning of agricultural stalks, and may assess a fine of USD 29 or less for farmers found violating the regulation. Burning is common in spite of the ban. Why the Inconsistency? -------------- ¶6. (SBU) Chengdu environmental officials tell us that air quality is measured daily from noon to noon, meaning that each recorded day of data actually includes the afternoon and evening hours from the previous day. Monitors placed at 38 sites around the city log pollution data and transmit it to the Chengdu Environmental Supervision Center (CESC). The results for the 24-hour period are then transmitted to Beijing and select officials across Chengdu. Some of these officials receive text messages each afternoon on their cell phones with the day's air quality before it is publicly available. ¶7. (SBU) CESC officials claimed recently that the discrepancy between some residents' perceptions of pollution and the recorded air quality resulted from brief periods of peak pollution that were not sustained over a 24-hour reporting period. They also explained that different parts of the city experience differing levels of pollution. On a day when air quality around several parts of the city appeared very poor, however, no monitoring station recorded higher than lightly polluted air. CESC said that monitoring stations around Chengdu registered API values ranging from 102-130 between noon on 17 May and noon the following day. The official average for the day was only 104, suggesting that the majority of stations around the city recorded at, or very close to, the bottom of the range. Reviewing the CESC data, we estimate that no more than one or two monitoring station reached 130, and few stations could have recorded a value higher than 110. ¶8. (SBU) Comment: We are unclear if the air quality monitors around the city record better than expected air quality because of the locations of the monitors or because CESC officials actually manipulate the data once they receive it. The one air monitoring station that Congenoff was able to visit in February did not seem like its location would provide artificially good air quality readings. Officials showed Congenoff the modem in the monitoring station that enabled the data transfer to CESC, but Congenoff has not had an opportunity to see where the data from all of the monitoring stations is aggregated. We do not know if any of the air quality monitoring stations have been moved over the last several years. BOUGHNER