Plum Rainy Season 2022

Plum Blossom Period 2022

July 8-15, 2022

Because each region has different temperatures, the time of entering and exiting the plum blossom in 2022 is also different. But the plum blossoms usually enter in mid-June and emerge in the first half of July, lasting about 20 days. But there are cases of late recruitment and late departure of plum blossoms. For example, in 2020, the plum rain area, the plum blossom into the plum early, out of the plum late, lasting a long time. That year, Zhejiang officially entered the plum rain season at the end of May, ten days earlier than before.

Generally, the 2022 rainy season will begin in early June and end in early July, lasting about twenty days. This year, it is expected to officially enter the plum season around June 10 and leave in mid-July in various parts of the country. The rainy season is coming, be sure to pay attention to things at home, check more, do not get moldy.

Jiangsu Taizhou_Plum Rain 2022

1. When will Jiangsu enter the plum season in 2022?

The 2022 rainy season in Jiangsu officially entered May on June 23rd. According to the latest press conference held by Jiangsu Meteorological Station and Nanjing Meteorological Station, it was announced that Nanjing officially entered the rainy season from June 23rd. In addition, areas south of the Huaihe River in Jiangsu Province are also expected to enter the plum season on June 23, so the rainy season in Jiangsu this year is Thursday, June 23rd.

1. Is Jiangsu a late bloomer this year?

It is a late plum blossom. Because the average perennial plum blossom day is June 19, this year's plum blossom day is June 23, a bit late. As the plum blossom belt swings from north to south, there is a lot of strong convective weather, with significant intermittent precipitation and phases of high temperature. Meanwhile, Huaibei will also enter a rainy period from June 23rd.

2. How is the weather in Jiangsu in May rainy season this year?

According to the latest introduction by the chief forecaster of Jiangsu Provincial Meteorological Station, the high temperature will continue in the province in the early part of this year's rainy season, and the strong convective weather will be more frequent. after June 24, the north-central area of Jiangsu Province is expected to have short-term heavy precipitation, thunderstorm winds and even hail, which will require extra precautions. Jiangsu is expected to have two significant rainfall process in the coming week, respectively, in the night of the 22nd to 24th and 27th to 28th. 23~26th, there are short-term heavy precipitation, thunderstorm winds, small hail and other strong convective weather. 22nd north-central, 23rd along the Jiangjiang River, Southern Jiangsu, 24th to 25th along the Huaihuai, Huaibei have more than 35 ℃ of high temperature weather.

3. How much plum rain is there in Jiangsu this year?

The average amount of plum rain is 200-260 millimeters. The average rainfall in Huaibei area during this period is 170-230 millimeters, which is more than normal.

2. When will the plum blossoms bloom in Jiangsu in 2022?

According to the latest introduction by the chief forecaster of Jiangsu Provincial Meteorological Station, the plum blossoms are expected to come out in mid-July 2022. The duration of the plum rain in Jiangsu in recent years is as follows:

1.2021 Jiangsu Provincial Meteorological Station issued a plum rain forecast, local areas south of the Huaihe River officially entered the plum rain on June 13th.

2.In 2020, the rainy season in Jiangsu started on June 9 and ended on July 21, and the rainy season lasted 43 days.

3. 2019 Jiangsu entered the Meiyu period from June 18th to July 21st. The total length of the Meiyu period is 33 days, which is longer than the normal Meiyu period of 23 to 24 days.

4. The rainy season in Jiangsu lasted 32 days in 2016.

Generally speaking, the 2022 rainy season in Jiangsu province officially entered the plum season on June 23 and usually comes out in July. According to the latest weather forecast for Jiangsu province, the plum blossoms will come out in early July this year.

Wuxi yellow plum days over 2022

The time of the 2022 plum rainy season occurs from late May to late June. Because every year the plum rain period occurs during the two festivals of Mangdiao and Xiaozhu, and this year Mangdiao is on June 6, while Xiaozhu is on July 7th.

So it is expected that China's Meiyu season in the middle and lower reaches of the Yangtze River will start from early June, and according to the time of entering Meiyu in previous years, it is not uniform across the region, and it will be a few days apart. Like 2021 Shanghai on June 10 into the plum; Jiangsu Suzhou on June 10 into the plum, south of the Huaihe River into the plum June 13 only into the plum.

Note.

2022 into the plum standard: five consecutive days the average temperature exceeds 22 ℃, there are four days for rainy days is considered to be into the plum. And according to the recent weather forecast for Shanghai, it has not officially entered the plum blossom yet, with lows still between 16-18 degrees.

The 2020 to 2022 epidemic trend

Big Data Epidemic Watch: Has the national epidemic peaked?

Tengking Macro Financial Big Picture Research

2022-12-2317:23 - From Beijing

Tengking Macro Snapshot

December 23, 2022

Big Data Epidemic Watch: Has the National Epidemic Peak Passed?

-Based on Tengjing AI High Frequency Simulation and Forecasting

Tengjing High Frequency and Macro Research Team

Highlights of this issue:

In response to the question of whether the forecasts are accurate or not, and whether the national outbreak has already peaked, we have added the daily metro passenger volume data of 28 cities for to assist in our judgment. The absence of a sample of non-internet users may have skewed the predictions.

Big data is not perfect, and the application of big data to make macroeconomic predictions is not perfect. We analyzed why Google Flu Trends failed. The reasons may include: the media's significant coverage of Google Flu Trends led to changes in people's search behavior, and users' search behavior in turn affects GFT's forecasts.

The current national epidemic may not have peaked yet, but the process of peaking may be ahead of schedule. With the help of subway passenger traffic data, we have determined that Beijing, Shijiazhuang, Wuhan, and Chongqing have already passed the peak of the epidemic, while Chengdu, Tianjin, Changsha, Nanjing, and Xi'an have not yet reached the peak.

I. Is the forecast accurate or not? Expectations and Reality Verify Each Other

In our previous report, "Big Data Epidemic Watch: Central Cities Take the Lead in the Peak", we analyzed and gave a prediction that the epidemics in some of the cities in Beijing and Hebei had reached an "inflection point", and that Chengdu, Kunming and other cities would reach their peaks one after the other. According to Baidu search index data, Beijing's Baidu search index for "fever" continues to decline, and the search index for "cough" peaked after "fever," which basically confirms our model's This basically confirms the prediction of our model. However, we also notice that the nationwide "fever" index peaked on December 17, 2022, does this mean that the national epidemic has peaked? If so, this data is different from the judgment of some epidemic prevention experts that the epidemic will peak around the Chinese New Year. Some experts believe the national epidemic may not have peaked yet, but the process has been shortened.

But according to ByteDance's "huge amount of math", ShakeYin's "fever" search index peaked on December 17th, but Headline's "fever" search index is still shaking. The search index of headline "fever" is still oscillating upward. In the circle of friends widely spread Zhihu "data emperor" prediction, December 20, 2022 around most provinces and cities have reached the peak of infection, then, many researchers want to confirm is standing in December 23, 2022, the nationwide single-day new infections have not reached the peak? Some people think that the prediction is accurate, and their own perception of the epidemic these days on the Internet is more consistent; some people think that it is not allowed, that relatives and friends around the Yang, and the prediction of the progress bar is less than halfway, individual perception and prediction of the results of the larger differences.

At the same time, we noticed that around December 16, 2022, the search index for "fever" in almost all cities and provinces in China had a pulse-like growth, and the subsequent daily data was never higher than the value on December 16, 2022, which was the same day as the one on December 16, 2022, and the daily data was never higher than the value on that day. The subsequent daily data was never higher than the value of the day of 16th. Does this mean that the toughest phase of the epidemic has passed? Data mining and modeling analysis of Baidu and Headline outbreak search engine data can provide an important reference for future trends in the epidemic. However, we understand that more data needs to be introduced in order to quantitatively evaluate the progress of the epidemic.

Because there is no authoritative data as a reference, the prediction of all types of epidemics is only based on intuition, reasoning or deduction with parameters of the model prediction, prediction is accurate, the lack of objective authority as a result of the comparison, so it is difficult to objectively measure the prediction of whether the accuracy of the prediction can only be achieved through the participation of this prediction of all viewers and readers through the microscopic data, the degree of diffusion of epidemics around the verification of prediction results, a The sequence of infection of different groups in a city, and the rhythm of the peak of infection in different cities will all have a different understanding of whether the prediction is accurate.

With the limitations of models, the applicability of logical assumptions, and the lack of authoritative data to validate them, are predictions not needed? Thomas Kuhn and Karl Popper engaged in one of the most influential confrontations of the 20th century over the concept of "philosophy of science". In their own profound ways, they both questioned the basic premises of science from a philosophical perspective. Kuhn's The Structure of Scientific Revolutions points out that even if the results predicted by an existing paradigm have counterexamples in reality, existing scientists do not consider their paradigm to be problematic; it is only when a new scientific paradigm emerges as an alternative to the existing paradigm, and when the number of counterexamples reaches a certain level, that the existing paradigm can be falsified and a scientific revolution can occur. The negation of the prediction process from a critical point of view is also the process of discovering new methods of prediction.

Philosopher Karl Popper, whom George Soros of the Quantum Fund promotes, is best known for arguing that science proceeds through "falsifiability" - one cannot prove that a hypothesis is correct, or even obtain evidence of the truth by induction! But if the hypothesis is wrong, it can be disproved. According to Popper, only theoretical systems that can be empirically falsified should be given the status of true science. Therefore, Popper advocated bold assumptions, and used falsification as a means of constant trial and error, and constant revision, rather than proposing hypotheses and then looking around for grounds to support his theories. "Falsification" is also a way of thinking that Soros has always promoted and practiced.

Second, the subway passenger volume as an important auxiliary observation index of the epidemic peak

So, we start from the epidemic, back to the economy, from the multi-dimensional verification of the peak of the epidemic. Subway passenger traffic is undoubtedly a good indicator to observe. The passenger traffic of a city with a subway is affected by a number of factors: 1) travel control, 2) willingness to travel, and 3) accessibility of the subway.

From the data point of view, Beijing, Shanghai as the country's two cities with the highest metro ownership, but also the highest average daily passenger volume of the two cities, the metro data higher reflect the high and low epidemic, while the metro passenger volume of the daily data published with a lag of 1-3 days, is still relatively timely, from the perspective of the collection of data, the metro data from the Internet of Things equipment automatically collected, the impact of human intervention is smaller, the data has sufficiently objective, the data is not a good indicator, the data has a lot to say, and the data has a lot to say. The data are sufficiently objective to be used as the second major observational variable of the epidemic.

Figure: Shanghai metro passenger traffic

▲Data source: Wind, Tengjing AI Economic Forecast

The above figure shows the Shanghai metro passenger traffic data from December 2019 to the present day, which is more evident in the Wuhan outbreak in early 2020, the Shanghai outbreak in April 2022, and the national outbreak in December 2022. As metro passenger traffic follows the principle of high Monday through Friday and low Saturday and Sunday, the daily data is somewhat informative and subsequently we can filter out short-term intra-day data fluctuations by comparing weekly averages.

Figure: Shanghai subway passenger traffic

▲Data source: Wind, Tengjing AI Economic Forecast

Comparing Beijing subway passenger traffic, it can also be seen that in April 2022, the Shanghai subway shut down for about 7 weeks, Beijing, although not shut down, but the average weekly subway passenger traffic from the last three years the daily 8 million down to less than 1 million. It is worth noting that Beijing subway passenger traffic after September 2022 was significantly lower than in Shanghai, partly because of the epidemic, but also because the Beijing subway required 72 hours of nucleic acid checks across the network, which was further shortened to 48 hours on November 24, and the policy was lifted as of December 5th.

Figure: Beijing subway passenger traffic

▲Data source: Wind, Tengjing AI Economic Forecast

Figure: 7-day moving average of subway passenger traffic in 10 major cities, with a high degree of consistency in synergism

▲Data source: Wind, Tengjing AI Economic Forecast

Based on this data, we believe that the peak of the outbreak in Beijing is over but that the national The overall epidemic peak has not peaked as shown by Baidu search index and headline index, but is in a period of rapid development. We have built a four-stage data model to assist in verifying whether each city has reached its peak. As shown in the chart below, Beijing, Wuhan, Chongqing, Shenyang, Shijiazhuang, Lanzhou, and Kunming metro passenger traffic has stabilized and is now in the fourth stage; Chengdu, Tianjin, Changchun, Zhengzhou, Guangzhou, Xiamen, Shenzhen, Xi'an, Shanghai, and Nanjing are still in the third stage in the process of reaching the peak. Since moving averages have the potential to bring about data lag, later on, we did a test with real data.

Figure: Epidemic spreading process

▲Data source: Tengjing AI Economic Forecast

Figure: Metro passenger traffic in some cities in China

Note: The top 10 cities are: Beijing, Shanghai, Guangzhou, Chengdu, Nanjing, Wuhan, Xi'an, Suzhou, Zhengzhou, Chongqing, and so on.

▲Data source: Wind, Tengjing AI Economic Forecasts

In the day-to-day progression of the epidemic, if there is a rebound in metro travel data on that day, you should look at two main figures, the first year-on-year, and the second year-on-year.

Based on daily data, Beijing subway travel, both ring and year-on-year, are in an upward phase, which is consistent with the top judgment, the other likely to top is Wuhan, Chongqing, Chengdu. And Shanghai, Guangzhou, Nanjing, Suzhou, Xi'an and other metro passenger traffic is still continuing to decline, which indicates that the epidemic is still in the process of reaching the peak.

Figure: Metro passenger traffic in selected cities in China

▲Data source: Wind, Tengjing AI Economic Forecast

Since the year-on-year decline in metro passenger traffic is severe, we judge that the epidemic is still in the process of reaching its peak in cities such as Shanghai, Guangzhou, Nanjing, Xi'an, Suzhou, Zhengzhou, etc., and that the epidemic is expected to have already passed its peak in Beijing, Wuhan, and Chongqing as they are turning positive year-on-year.

Figure: 28 cities metro passenger traffic and weekly year-on-year

▲Data source: Wind, Tengjing AI Economic Forecast

Three, how do expectations interact with reality?

There is a lot of experience after the deregulation of the epidemic, whether it is the rhythm of the epidemic topping out, the impact on consumption, labor participation rate, there are more countries to refer to. This certainly gives us something to look forward to, the 1.4 billion population of the liberalization and the medium-sized population of countries to liberalize and there is a difference. Domestic infectious disease experts also said in various media around the Spring Festival, the epidemic peaked in the first quarter of next year, and so on, releasing such a future top signal. But from Beijing and most of the city's perception, the epidemic seems to peak earlier than we know, so where in the end will go wrong?

Policy Indicator Failure: Goodhart's Law

When most Internet players know that the Baidu search index is an indirect proxy for the epidemic, it may not be accurate, and to some extent it is Goodhart's Law for the epidemic. Goodhart's Law is a phrase from British economist Charles Goodhart, which refers to the idea that when a policy becomes a goal, it will no longer be a good policy. One interpretation of this is that once a social or economic indicator becomes a stated goal to guide macro policy making, the indicator loses its original informational value.

There is no doubt that the "Baidu Epidemic Index" is probably still valid when most people are unaware of its significance, and the internal logic is that the search volume data indirectly reflects the spontaneous online search behavior of most residents, and that "fever" searches are, to a certain extent, the most important factor in the development of the disease. "Fever" search to a certain extent and positive symptoms is the same thing. But with both official and self-published media reporting, this indicator triggers more searches that have nothing to do with the outbreak itself, but rather the effect of Internet traffic.

Data contamination may be caused by skewed search behavior

We compared subway ridership in Shijiazhuang, Lanzhou, Beijing, Wuhan, Chongqing, Shenyang, Kunming, Chengdu, and Tianjin, and found that they all experienced a pattern of upward movement in response to easing of the policy, downward movement in response to the outbreak, and then upward movement in response to the epidemic's peak. At present, most cities are still in the epidemic climbing passenger traffic down this stage, the peak of the national epidemic has not come, and Baidu index gives the "fever" search index has peaked, we judge that after December 16 and Baidu "fever" search index may be abnormal. The core logic is that on December 16th, all cities across the country saw a spike and then a drop, and it's very likely that this kind of factor affecting all cities at the same time wasn't caused by a virus spreading at a certain rate, but rather by data "pollution" caused by other factors.

Missing sample: non-internet users aged 60 and above

We know that the Baidu Index, Headline Index, and Micro-Index are data products based on data mining and analysis of massive internet user behavioral data, so non-internet user behavioral data is naturally excluded from the research sample.

The 50th Statistical Report on the Development of the Internet in China, released by the China Internet Information Center on August 31, 2022, shows that as of June 2022, the size of China's non-Internet users was 362 million, which is not a small base. From a regional point of view, China's non-Internet users are still predominantly in rural areas, and the proportion of non-Internet users in rural areas is 41.2%. From the age point of view, the elderly group aged 60 and above is the main group of non-Internet users. Accordingly, it can be seen that non-Internet users are geographically distributed mainly in rural areas, and age is dominated by the elderly group aged 60 and above.

The lack of search behavior of this non-Internet user group with a small base has led to the search results that should have appeared outside the sample, resulting in the underestimation of the search index for diseases such as "fever". According to the Centers for Disease Control and Prevention, the risk of severe COVID-19 increases with age, disability, and underlying disease. During the late omicron period, most in-hospital deaths occurred among adults aged ≥65 years and those with three or more underlying diseases.

Figure: daily confirmed cases of COVID-19 in countries and territories around the world

Note: due to limited testing, the number of confirmed cases is lower than the true number of infections, data as of December 21, 2022

▲Data sources: johns hopkins university CSSE COVID-19 database, ourworldindata. org, Tengking AI Economic Forecasts

Figure: daily confirmed COVID-19 cases in all regions of the world

Note: due to limited testing, the number of confirmed cases is lower than the true number of infections, and the data is as of December 21, 2022

▲Data source: the Johns Hopkins University CSSECOVID-19 database, ourworldindata. ourworldindata.org, Tengking AI Economic Forecasts

Big data isn't perfect, so why is Google Flu Trends failing?

As early as 1980, futurist Alwin Toffler proposed the concept of "big data" in his book The Third Wave. Since ancient times, prediction has been a much-anticipated ability, and big data prediction is the most central application of data, the logic of which is that every unconventional change must have signs beforehand, and every thing has a trace, and if you find the pattern between the signs and the changes, you can make a prediction.

The use of big data methods and techniques for macroeconomic research and analysis has international precedents. In the vision of big data analysis, it is not only to figure out the macro-statistical laws, but also to figure out the fine structure in the macro-data. Based on the perspective of research, the era of big data provides powerful support for macroeconomic analysis and is changing the macroeconomic research paradigm.

Mainstream financial institutions, such as central banks, have developed and adopted real-time forecasting models to track changes in the state of the economy in real time, and to find reliable sources of information before they are overwhelmed by large amounts of socialized information, so as to dynamically adjust their expectations of economic indicators. These include the New York Fed's Nowcasting model, the WEI model, the Atlanta Fed's GDPNow model, and the Bank of England's MIDAS model.

According to Prof. Didier Sornette's "Dragon King" theory, there are two conditions for the occurrence of extreme events: system consistency and synergy. When the consistency of the system is very strong, the black swan type of extreme events are easy to happen. When the system's consistency and synergy are strengthened at the same time, a more extreme "Dragon King" event beyond the "Black Swan" will occur.

Black swans and dragon kings are not isolated events, but rather a series of strongly related events that reflect the power of positive feedback. When can the stock market be predicted? The key lies in the degree of correlation between the stock market changes before and after.

In 2008, Google launched the GoogleFluTrends system, which is motivated by the ability to detect disease activity early and respond quickly to reduce the impact of seasonal and pandemic influenza, by analyzing a large number of Google search queries collected to reveal the presence of influenza-like illnesses in the population. The logic and idea is actually quite simple and intuitive - if you're sick, you're likely to search the search engine to find information such as how to treat it. Google decided it wanted to track those searches and use that data to try and predict flu epidemics, even before medical organizations like the Centers for Disease Control could do so.

Google Flu Trends became famous in 2009 when it successfully predicted the spread of the H1N1 flu in the U.S. through the massive amount of search data Google had accumulated. It was reported that Google Flu Trends was able to predict regional influenza outbreaks 10 days before the Centers for Disease Control and Prevention reported the outbreak.GFT's ability to predict such outbreaks is obviously of great social significance, and it can win a head start in controlling infectious disease epidemics in advance for the whole society.

So Google created a nifty equation on its Web site to figure out just how many people have contracted the flu. The simple-to-understand logic of the data goes like this: people's locations + flu-related search queries on Google + some very clever algorithms = the number of people with the flu in the United States.

Linear models were used to calculate the log odds of an influenza-like illness visit and the log odds of a related search query:

P is the percentage of doctor visits, Q is the fraction of queries related to ILI calculated in the previous steps. β0 is the intercept, β1 is the coefficient, and ε is the error term.

Google Flu Trends has been shown not to be consistently accurate, particularly between 2011 and 2013, when it overestimated relative influenza incidence and predicted twice as many visits as recorded by the CDC for one time period during the 2012-13 flu season.A 2013 article published in Nature claimed that Google Flu Trends overestimated flu cases by by about 50 percent.

As you can see, applying big data to make macroeconomic predictions is not perfect. According to economist and author Tim Harford, "the failure of Google Flu Trends highlights the dangers of unfettered empiricism." One explanation for GFT's failure is that the news is full of

Figure: Google Flu Trends ILI estimates compared to CDC estimates

▲Data source: ImprovingGoogleFluTrendsEstimatesfortheUnitedStatesthroughTransformation,LeahJMartin,BiyingXu,YutakaYasui,Tengjing AI Economic Forecasting

In 2013, Google In 2013, Google adjusted its algorithm and responded that the "culprit" for the bias was a change in people's search behavior as a result of the media's heavy coverage of the GFT, which did not appear to take into account the introduction of specialized healthcare data and experts' experience, and did not "cleanse" and "decontaminate" user search data. It also did not "clean" or "denoise" user search data. After 2011, Google introduced "Recommended Related Search Terms", which is the familiar search term model we are familiar with today. Researchers analyzed that these adjustments may have artificially pushed up some search indices and led to an overestimation of the prevalence of the epidemic. For example, when a user searches for "fever", Google will also give "sore throat and fever", "how to treat a sore throat" and other related recommendation words, then users may be out of curiosity! Users may click on the keywords out of curiosity, resulting in the phenomenon that the keywords used by users are not intended by users, thus affecting the accuracy of GFT search data. The user's search behavior in turn affects GFT's prediction results. In the noisy world of search engines, which is full of media reports and users' subjective information, there is also the paradox of "prediction is interference". A similar situation is likely to occur in the domestic search engine index, which is an explanation for the difference in expectations based on our experience with GFT.

Figure: GFT's "fever" associated search terms

▲Data source: GFT, Tengjing AI Economic Forecast

References

[1]CNNIC: 50th Statistical Report on the Development Status of the Internet in China

[2]< /p>

[3]AdjeiS,HongK,MolinariNM,etal.MortalityRiskAmongPatientsHospitalizedPrimarilyforCOVID- 19DuringtheOmicronandDeltaVariantPandemicPeriods-UnitedStates,April2020_June2022.MMWRMorbMortalWklyRep2022;71:1182_ 1189.DOI:

[4]

[5]

[6]Lazer,D.,R.Kennedy,G.King,andA.Vespignani.2014. "TheParableofGoogleFlu. TheParableofGoogleFlu: TrapsinBigDataAnalysis. "Science343:1203_1205.

For more important research results, please pay attention to the public number "Tengjing AI economic prediction".

Search

Tianjin infection peak prediction

How many new crown patients died in Tianjin

Total number of deaths in the national epidemic

How many people have died in China's outbreak

Central government issued the latest policy of the epidemic

National outbreak update

What months are the rainy seasons in Xi'an

in 2022?

Xi'an is one of the more unique cities, it has various cultural heritage, and a variety of food and snacks, which is loved by people. In recent times, it has been raining in Xi'an, and it has been in rainy weather, which is a relatively normal phenomenon, and it is mainly caused by the subtropical high pressure, global warming, and geographic location.

2021Why Xi'an likes to rain in September

1.Subtropical high pressure

In September, Xi'an rained for more than ten days. From the meteorological data of past years, it is normal for Xi'an to have a lot of rain in September. In fact, the likelihood of the next ten and a half days is relatively high.

Xi'an has a warm temperate semi-humid continental monsoon climate with moderate rainfall and four distinct seasons. Winter is cold, windy, foggy, with little rain and snow; spring is warm, dry, windy and changeable; summer is hot and rainy, with prominent summer droughts and thunderstorms and high winds; and fall is cool. The annual precipitation is 500-750mm, mainly in summer and fall; Xi'an has long been in the northwestern part of the subtropical high pressure in summer and fall, and southwest and northeast winds prevail in winter.

The subtropical high pressure occupies the Pacific Ocean in the northern hemisphere in winter. As the point of direct sunlight moves northward, the subtropical high pressure also gradually moves northward. The northwestern edge of the subtropical high pressure is easy to combine with cold air to form precipitation. However, affected by topography, subtropical high pressure strength and other factors, spring precipitation is mainly concentrated in east and south China, which also leads to a precipitation peak in Xi'an around May. In summer, Xi'an is controlled by the subtropical high pressure, with more short-term rainstorms. When fall comes, the northwestern edge of the subtropical high pressure passes through Xi'an again as it retreats to the south, leading to continuous precipitation in Xi'an in September.

2. Global warming

The effects of global warming are complex. At present, the overall reflection of rainfall is a northward shift of the rainfall belt, but this northward shift is not merely a leveling off. Its scale and extent are locally specific. For example, against the background of a gradual increase in global temperatures and a northward shift of the rainfall belt, Shaanxi Province experienced a gradual decrease in precipitation from the 1990s to the beginning of the new century.

3. Geographic location

In fact, the Guanzhong Basin, in which Xi'an is located, is not rich in water systems and has a relatively small water area, making it difficult to create large amounts of localized thermal convection. To the south of the basin are the Qinling Mountains, the highest mountain range in the east. For Sichuan, the southwesterly flow of the subtropical high pressure in the northwestern Pacific Ocean transports warm, humid air from the Indian Ocean to the Sichuan Basin, where it meets cooler air on the northern Tibetan Plateau, creating a persistent fall rainfall in western China in September and October. However, the presence of the Qinling Mountains makes it difficult for much of the warm, moist air to enter the Guanzhong Basin as it climbs the southern side of the mountains to form topographic rains, which directly contributes to the two distinctly different wet and dry climates of Guanzhong and Hanzhong.

When is the rainy season in Xi'an

The rainy season in Xi'an is July, August and September. Xi'an has two distinct peaks of precipitation, in July and September. The average annual precipitation in Xi'an is 558-750mm, increasing from north to south. It changes every year.

In September, southern China, the area around the Tropic of Cancer, is far from cool, with warmer air still hovering there, waiting for cold currents from deeper in Eurasia to chase them away.

Not only in southern China, but also in the subtropics of South Asia and the Middle East, they are waiting for the same result. In addition, because both subtropical high pressures are off the coast, a lot of water vapor is also steaming up, but because of the heat, not much of it is condensing into rain.

From September to October, the subtropical high pressure moves southward and the rain belt returns to western China. There are said to be cloudy and rainy weather. This continuous fall rain also has a scientific name, called "autumn rain in western China" and "autumn rain" in Shaanxi. It is common in some areas of western China and usually occurs in Xi'an in September. Under the influence of subtropical high pressure in the south, the weather usually lasts about two to three weeks.

How to dry clothes faster on a rainy day

1. Paper towel press

After washing your clothes, no matter how hard you wring your clothes, there is always a lot of water on them. You can iron your clothes with paper towels. Paper towels are very absorbent. More paper towels can make the water on the clothes dry.

2. Wringing out the towel

We use a dry towel to help wring out. First wrap the wet clothes in a dry towel and then wring it out vigorously. At this point, the water on the clothes will be absorbed by the towel. It is best to choose a towel that is highly absorbent.

3. Add the dry towel and shake well

We can also dry it in the washing machine. We can dry it in the washing machine once and then in the second