The Potential Impact of Global Warming on China's Mortality from Different Categories of Diseases

(Not for publishing and citation)

Guanri Tan

***Tables and figures not available at this time***

Deaths in 1979-1989 were classified into ten categories: INF (infectious diseases), CAN (cancer), END (diseases concerning endocrine and blood), PSY (psychosis), CIR (circulatory diseases), RES (respiratory diseases), DIG (diseases in digestive system), GYN (diseases in gynecology, obstetrics and newborns), OTH (other diseases not included in the above categories, and excluding deaths in accidents) and ACC (accidents). TOT stands for total mortality of all causes.
Since the categorization of diseases did not strictly follow the international standard, and might vary from city to city, so the uses of the figures should be with caution.

The top cause of death is CIR, accounting for 31% or 24% of the total of the two cities respectively (first number is for Shanghai and the second is for Guangzhou hereafter). The following causes are CAN, RES and DIG. These four account for up to 77% or 66% of the total death. OTH is the sum of deaths from many uncommon diseases, accounting for a rather big portion of the mortality. Although the two cities are far apart from each other, the percentages for the same diseases for the two cities are very similar. This implies that the relative severity of the diseases for these two cities may represent conditions over very wide areas in eastern and southern China.
Mortality from most categories of diseases shows great seasonal change. This indicates that climate imposes great influences upon morbidity and mortality.

The table shows that RES has the biggest ratio of winter mortality to summer one. This is due to the fact that cold air stimulates respiratory system, and there is always heavy air pollution in winter. So patients of RES suffer from bad environment in winter. Next to RES is CIR. Low temperatures in winter make the vascular system shrink, increasing blood pressure and enhancing vascular spasm. Actually, respiratory system mutually affects circulatory system. It's not surprising that they have similar annual variation. Many studies show that high mortality always occurs around the passage of cold fronts. The values of ratio for Shanghai are generally greater tnan those for Guangzhou. This means that human morbidity and mortality in Shanghai is more greatly influenced by season than in Guangzhou. This may be partly explained by the climatic feature that annual amplitude of temperature in Shanghai (26.2-4.9=21.9 degrees) is much bigger than in Guangzhou (28.2-14.2=14.0 degrees). The ratios for CAN are close to unity indicating that this disease is least affected by climate, since CAN is chronic change in organs in human body which is not due to a quick response to outside environmental elements.

The determination of threshold temperature followed the procedure we used before, but two more conditions to detect the most significant threshold were employed: 1) the tangent of the regression line of the higher group must be larger than that of the lower group. It is our experience that hot weather imposes bigger impact on mortality than cool weather does; 2) the difference between mean mortalities of the two groups must pass the t-test at 0.01 significance level.

From this set of new criteria, Guangzhou's thresholds for total mortality and for deaths from most diseases are generally one degree or two higher than Shanghai's counterparts. Threshold for CAN is higher than for TOT in both cities. Differences among most thresholds generally are not greater than one degree. Diseases that induce greater mortality on hot days than on normal days are RES, END, OTH and ACC. Diseases that are less influenced by temperature are CAN, DIG and CIR. The disease that is least affected by temperature is INF. For the simplicity of making Table 4, days with temperature equal to and higher than 34 degrees are refered to as hot-day (with different termal levels). The remaining days are called non-hotdays. Table 4 gives differences of deaths on various thermal days.

Diseases that significantly increase mortality when temperature gets higher are OTH, ACC and RES. Mortality from CAN doesn't change much with temperature.
Stepwise Regression Analysis picked out the main meteorological elements which affect mortality most on hot days.

The most influential element is temperature, especially maximum temperature or noon-time temperature. The higher the temperature, the more the mortality from all categories of diseases. Another rather important element is wind. The stronger the southern wind, the lower the mortality. There is a negative correlation between dew-point temperature and some diseases in Shanghai. At a glance, the phenomena that less deaths occur during humid weather seems not to consistent with common knowledge. The cause of this is that many dying patients died under the weather controlled by a subtropical high characterized by hot, dry and calm -- the most offensive summer weather in eastern and southern China.

The data in the table shows that there is a positive correlation between consecutive hot days and deaths from many diseases in Shanghai. This means that Shanghai people are not accustomed to heat waves. On the other hand, a negative correlation between 'date' and deaths for Guangzhou means that Guangzhou people would adapt to heat stress after living in hot weather for weeks.

More than 56% of the variance of deaths of TOT and OTH for Shanghai could be explained by weather variables (statistics not published here). More than 41% of the variance of CIR, RES and ACC is explainable by weather. These indicate that stressful weather is an important factor to Shanghai mortality. On the other hand, Guangzhou people are less impacted by weather, only 25% of the variance of TOT is related to weather. Even a less percentage (no more than 20%) of the variance of many categories of deaths is explained by meteorological element. However, weather is still a factor, accounting for deaths of one fifth of the total , and should not be neglected.

Winter conditions are quite different than summer. The most sensitive elemt to mortality is temperature. Mortality increases slowly with a decrease in temperature. Threshold temperatures were not detected for five categories of death. Thresholds for the other categories were detected, but their effects were very weak. Therefore, threshold temperature is not an feature for winter. Total deaths and deaths from CIR and RES respond to low temperature with a two-day lag. A few categories have a one-day lag. Few categories have a three-day lag. The delayed response of death to winter severe weather is partly due to the fact that there are more and better measures against cold than heat in human life.

Table 6 shows that mortality increases on the day(s) (1-3 days) following the invasion of a cold air mass with a chilly night, dry air, a strong north-western wind and less sunshine. Cold air invasions in late winter bring more mortality than those in early winter.
The disease in Shanghai mostly affected by weather is OTH, 19% of whose variance could be explained by meteorological variables (statistics not published here). Less than 10% of the variance of the deaths from the remaining categories is explained by weather variables. But there is more weather-related mortality in Guangzhou than in Shanghai. 27% of the variance of TOT, 20% of OTH and 16% of CIR are explained by weather. The explained variance of the remaining categories are less than 10%.

Table 4 shows that with lowering temperature, only deaths from END and ACC decrease, but all other categories increase. Among them, OTH, RES and INF increase faster. The increment of temperature in Table 7 is five-degrees, while in Table 4, one-degree. This shows that change in mortality in winter is much less than on summer hot-days.

Two scenarios were used. The slight warming scenario is the medium value of the scenarios recommended by China's Climate Change Committee. Increase in summer average temperature for both Shanghai and Guangzhou is 0.6 degree only. In winter, 0.7 and 0.9 degree for these cities respectively. The severe warming one is the GFDL scenario. Increase in summer temperature is 3.94 for Shanghai and 3.47 for Guangzhou. In winter, 4.55 and 3.44 respectively. Other meteorological elements are much less important so the future values remain as the present ones.
Comparison of mortality in Shanghai and Guangzhou implies that Shanghai people will gradually acclimatize to a warmer climate in the long run. So a partial acclimatization in summer is mostly possible. To obtain the summer Shanghai mortality under partial acclimatization, assume that: 1) the threshold temperature for Shanghai approaches Guangzhou's (34 degrees); 2) people get used to a heat wave-- term of 'consecutive hot day' in the equation is deleted; 3) referring to the findings for the U.S.A., in a severe warming scenario, mortality in partial acclimatization is equivalent to 56% of that in a situation without acclimatization. There is no problem of unacclimatization with winter mortality.

It has to be kept in mind that the percentages for categories should not simply be compared to one another, since they do not reflect the absolute change (amount) in death. If climate changes slightly as China's scenario predicted, deaths from most categories in summer for Shanghai will increase by around 1%, and will almost offset the decrease in winter mortality, so net change in annual mortality rate will be little. The only significant change is that the decrease in deaths from OTH in winter surpasses the increase in summer. The percentage changes for Guangzhou are larger than Shanghai. Percentage increments for summer for Guangzhou are bigger than the percentage deduction for winter, but the annual absolute number of deaths will not change much since the base amount of death in winter is greater than in summer.

If climate warming is as predicted by GFDL, the number of hot days in summer for Shanghai will increase up to 54 days in an average summer, which is 42 days more than in the present. Hot days in Guangzhou will number 79, which is 55 days more than the present. Not surprisingly, heat-related mortality will increase dramatically. Percentage increments of deaths from CIR, RES, OTH and ACC will be up to two digits. Deaths from OTH will be especially higher than present, increasing by 30%. Winter mortality will obviously drop, but it can not offset the rise in summer mortality. Annual mortality will significantly go up. The mentioned estimates of future mortality are inferred from the acute impact of short term weather. As imagined, morbidity and corresponding mortality will also be affected by climate long-term warming, but the climate effects on morbidity are beyond the scope of this study. However, a qualitative estimate of impact of climate warming on China's mortality could be made: summer will be a terrible time; spring and fall will be not much different from what they presently are; winter will be a pleasent time for health.