Meteorology is a science, precision monitoring is the foundation, precision forecasting is the core, and fine service is the goal. Let’s talk about the past and present of the weather forecast today! (Comrades who are interested in weather monitoring, please move forward: How precise is weather monitoring! )
The weather forecast has always attracted people's attention and profoundly affects human production and life. From the initial metaphysics, weather forecasting was judged purely by life experience. Later, when the traditional weather map was born, weather forecasting began to become an applied science, and then to the current numerical weather forecast, the forecast accuracy rate has greatly improved. The evolution of science:
1. The development of early weather forecasting
Three thousand years ago, in my country’s Oracle, there were records of actual weather conditions, including wind, cloud, rainbow, rain, snow, frost, haze, tornado, thunderstorm, etc. Ancient people were greatly affected by the unpredictable weather. The weather was often related to divination and rain-seeking activities, which belonged to metaphysics.
For thousands of years, sailors, fishermen, farmers and hunters have used cloud, wind, and astronomical observations to predict the weather and explore weather forecasts. As early as 2000 years ago, our country summarized the twenty-four solar terms, which mainly guided agricultural production and sacrificial activities. The weather proverbs that have been passed down to this day can also reflect the early people's attention to weather forecasts, such as "Do not go out in the morning, and travel thousands of miles in the sunset", "The sky hooks the clouds, the ground rains". In "The Romance of the Three Kingdoms," Zhuge Liang used the east wind to burn Cao Jun’s warship and the foggy day grass boat to borrow arrows, which is the application of short-term weather forecasts in the military.
The development of weather forecasting is inseparable from quantitative observations. With the development of ancient science and technology, various weather monitoring instruments continue to emerge, including thermometers and rain gauges invented by Galileo and his disciples, barometers invented by Torricelli, and Kote Invented hygrometer, etc.
With the invention of meteorological instruments and the establishment of meteorological observation networks, as well as the development of fluid dynamics theory, weather forecasting has gradually shifted to the direction of applied science.
2. The Birth of Modern Weather Forecast
The invention of the telegraph has provided conditions for the rapid transmission of meteorological observation data and the concentration of information in various places, making it possible to draw real-time weather maps. The birth of weather maps is the symbol of the birth of modern weather forecasting.
This is the first weather map made by the Chinese in 1915. Compared with the West, the arrival of this picture has been a hundred years late.
The story about the development of war-driven weather forecasting is well-known. During World War I, the British government, which was the absolute leader in weather, stopped providing meteorological data to the Norwegian government. The Norwegian government started with only 8 observation stations. Meteorological monitoring, meteorological talents, theoretical research and other directions have made great efforts to achieve the famous Norwegian school that followed. Open any of the current "Principles of Weather" books, circulation theory, cyclone structure, cyclone life history, cold and warm fronts, air masses, etc. are all from the Norwegian school, and Pieknis is the core figure of the Norwegian school. They have made outstanding contributions to weather forecasting. Looking at China today, a technological blockade can also become the best opportunity for self-development.
The era of relying on weather theory and forecasters' experience to manually draw and analyze weather maps, and then make weather forecasts has ushered in great development with the dissemination of radio, television, newspapers and other media. At present, most forecasters have gone through this stage. They analyze, discuss, and forecast short-term weather and forecast success. They are helpless in the face of medium- and long-term weather forecasts. And then the advent of numerical weather forecasting equipped forecasters with the darkest technological weapons, and they began to smile in the weather forecast. Relying on numerical weather forecasting, the current weather business can achieve seamless weather forecasts: short-term (0-a few hours), short-term (1-3 days), medium-term (4-9 days) and extended periods (10-30 days).
3. The advent of numerical weather forecasting
Numerical weather forecasting is to solve the atmospheric motion equations by numerical methods given initial and boundary conditions, so as to predict the atmospheric state at a future time from the atmospheric state at the known initial moment.
In 1904, Pieknis proposed the idea of numerical forecasting. From 1916 to 1918, Richardson made numerical forecasting attempts. You must know that there was no computer at that time. Richardson designed a computer factory and required 64,000 people to calculate at the same time in the factory. It ended in failure.
After that, after the shocking operations of many meteorologists like Rossby, Charlie, and von Neumann, the first number of weather forecasts (NWP) appeared in Stockholm, Sweden in 1954, and it is still unavailable in actual business. During the period, it experienced the rapid development of meteorological satellites. In the 1980s, the European Center for Mid-range Weather Forecast (ECMWF) began to produce global forecasts every day. In the following decades, EC will lead the world's numerical weather forecasting centers to move forward.
The business operation of EC's first four-dimensional variational (4D-Var) data assimilation method is a milestone in global operational numerical weather prediction. In the following decades, thanks to the continuous optimization of four-dimensional variational data assimilation, as well as the development of meteorological satellite observations and supercomputing equipment, the number of effective forecast days for weather forecasts increased by about one day every ten years. The convergence of the northern and southern hemisphere curves after 1999 is due to the breakthrough brought by the use of variational methods to process satellite data.
The operational operation of ensemble forecasts is also a highlight in the history of numerical weather forecasting. Everyone knows the "butterfly effect", that is, a butterfly in the tropical rainforest of the Amazon River Basin in South America, occasionally flapping its wings, can cause a tornado in Texas in the United States two weeks later. Meteorologist Lorenz pointed out in 1963 that the weather system is a chaotic system, and a small error will make a huge difference over time, that is, the error of the initial conditions makes the weather forecast more inaccurate in the future.
Ensemble forecasting makes up for the chaotic nature of the weather system. Since the true initial value cannot be obtained, a lot of initial disturbance values are used for simulation, and many different forecasts are obtained. Some members may also have different forecast results. However, the probability forecasts obtained by ensemble forecasts provide the weather forecast users with important and scientific information.
At present, global numerical weather forecasting is in a super-strong situation, and EC is the well-deserved boss.
The level of world numerical weather forecasting can be roughly divided into three camps
The first camp: European Center for Medium-Term Weather Forecast (ECMWF)
The second camp: United Kingdom Meteorological Office (UKMO), National Center for Environmental Prediction (NCEP), Japan Meteorological Agency (JMA), Canadian Environment Agency (CMC), Meteo France (Meteo France), German Meteorological Agency (DWD)
The third camp: China (CMA), Brazil (CPTEC), Australia (CAWCR), South Korea (KMA)
4. The trend of future weather forecast
With the continuous development of high-performance computing, observing systems, and meteorological theory science, future development paths for weather forecasting include but are not limited to:
1. Develop a coupled atmosphere-land-ocean-sea ice earth system model; the temporal and spatial resolution continues to improve, and global kilometer-level numerical weather prediction will become possible; the integrated weather-climate model will gradually become the mainstream. Of course, these seemingly clear development directions still face great challenges, including physical process parameterization, cloud radiation processes, aggregation methods, and breakthroughs in various meteorological theories.
2. The application of artificial intelligence in weather forecasting. Weather forecasting and artificial intelligence are naturally coupled. The outstanding big data processing capabilities of artificial intelligence have become an important tool for further accurate weather forecasting. Foreign countries have realized the research and development of data assimilation algorithms based on deep neural networks and meteorological satellite observation data. Within a certain tolerance of accuracy, the computational efficiency of artificial intelligence methods can be greatly improved compared with traditional methods. In recent years, the European Medium-term Weather Forecast Center has comprehensively evaluated the application potential of artificial intelligence technology in various technical aspects of weather forecasting numerical models, and has given optimistic expectations for the application of artificial intelligence, and has already carried out some aspects such as physical process parameterization. Carry out technical trials in China.
references:
1. Professor Qian Weihong, Peking University "The Development Process of Weather Forecasting"
2. European Center for Medium-Term Weather Forecast www.ecmwf.int
3. [Forecasting the weather, artificial intelligence is better than humans? ], Science and Technology Daily http://www.cma.gov.cn/2011xwzx/2011xmtjj/202007/t20200730_559576.html
4. [The Quiet Revolution of Numerical Weather Prediction] The Quiet Revolution of Numerical Weather Prediction http://www.nature.com/nature/journal/v525/n7567/full/nature14956.html