China is about to detonate meteorological data

20 meteorological data increased thousands of times times

As with the process of the world's largest data age, the volume of meteorological data has doubled.

In the last century http://www.aliyun.com/zixun/aggregation/16333.html ">90 years and before, the meteorological data mostly confined to the ground and high-altitude observation." At that time, more than 2000 ground stations collected meteorological information on an hourly basis, and more than 120 high-altitude stations observed no more than 4 times a day. Not too much from the amount of data, even considering satellite and radar data, the overall daily increment is limited to the GB level.

At present, there are about 40,000 terrestrial observatories, observed every 10 minutes, and will be encrypted to minutes in the future, at least 20 times times more in space density, 60 times times more frequently, and 1200 times times the total amount of ground and altitude observation information.

And these accounted for only 30% of the entire meteorological data, radar, satellite and numerical forecast data accounted for 70%. At present, the annual meteorological data is close to PB level (1000GB=1TB,1000TB=1PB).

This is also the embodiment of large data laws, that is, the relatively simple operation of large data is always more accurate than the results of complex operations on small data. The greater the amount of information there is, the better the prediction can be.

Shen Wenhai, Deputy chief engineer of the National Meteorological Information Center, summed up the large data features of the meteorological Department: to a certain extent, the meteorological department has been using large data methods, such as climate factors "positive correlation" "remote correlation" and some statistical methods.

The industry-defined "big Data" feature, with the exception of large capacity, multiple types, high speed growth, and a scarcity of value, that is, finding the valuable data needed in a large amount of data is like looking for a needle in a haystack, and it needs to be calculated by the corresponding mathematical model. In the weather industry, although the amount of data has gradually become larger, but each data has a specific value, so the meteorological industry data does not fully conform to the industry defined "large data" characteristics.

NI, academician of the Chinese Academy of Sciences, so far, "big data" mainly used in the social sciences sector, such as government, public health, social Security and other departments, and the use of natural science in the case of few cases.

Meteorological services Activate data

How to use the massive meteorological data? This is an issue that needs to be considered in the big data age. As far as the present situation is concerned, the data are used more fully in weather forecast and climate prediction diagnosis, while in the Meteorological service field, a large number of live observation data are often put on hold.

The current data meteorological service is mainly based on single element single site form. This means that the meteorological services that people receive are just weather conditions at the perimeter meteorological site, and there is always a delay.

To this end, the researchers are introducing the international advanced Spatial Data Fusion numerical model method, the surrounding sites of data and other sensors in the data fusion into the model, the entire region to reverse the weather. From the test results, the operation speed reached the minute level, the small area can reach the second level.

"These work are on the basis of large data can be carried out, no matter how advanced the model, no massive data entry, can not achieve very good results." "China Meteorological Bureau Public Meteorological Service Center senior engineer Tangqioghong said."

To the delight of researchers, in the big data age, data is not simply about people posting on the Internet. The world's industrial equipment, automobiles, electric meters have countless digital sensors, at any time to measure and pass on the location, temperature, humidity and even the chemical changes in the air. It can be imagined that this information can be used by the meteorological Department.

What is the weather service like in the big data age? Tangqioghong that in the visible future, the integration of geographical information, socio-economic data of the meteorological services, can let people know what can happen at any time and place, such as whether the wind will blow the door billboards, a high-speed junction in front of the rain, will not happen to flash floods.