3:30 PM - 5:00 PM
[O08-P43] Forecast Temperature Of Nagano
~Using my own researched data~
1 Research motive
The changing weather is very interesting. I observe the sky every day and make simple weather forecasts. Among them, I noticed that the temperature forecast announced by the Nagano Local Meteorological Observatory (hereinafter: Japan Meteorological Agency: JMA) is different from the actual temperature. Is it possible to predict more accurately?
2 Current status (analysis of previous research)
(1) About the JMA forecast
JMA uses a supercomputer to make "numerical weather prediction". It calculates various factors such as weather and temperature. The calculation results are revised and announced by local staff.
(2) Accuracy of JMA forecast
In July 2021, as a result of investigating the expected temperature and the actual temperature of the Japan Meteorological Agency in Nagano, there was a difference of 1 ° C or more in about one-third.
(3) Research content
In order to forecast the minimum and maximum temperatures more accurately, I decided to make a temperature forecast myself. This time, I will use the data I collected to forecast the temperature and study whether the forecast accuracy will improve. We will also compare it with the Japan Meteorological Agency.
3-1 [Research 1] method
I recorded The weather, wind speed, and temperature from AMeDAS data every day, and investigated the tendency of temperature changes due to atmospheric phenomena. Enter data into a spreadsheet and categorize it to find regularity. Also, organize the weather charts into slides and save them.
3-2 Results of [Research 1]
The difference from the highest temperature to the lowest temperature on the previous day had regularity as shown in [* Fig. 1]. More than half of the atmospheric phenomena of each item are within 1 ° C around a certain value. It is a little wider on cloudy days. There was certainly regularity due to the wind speed [Fig. 2], and the change in temperature was smaller when the wind was stronger. It was found that the width in the graph depends on the wind speed.
4-1 [Research 2] method
Apply the trends and regularities obtained in [Research 1] to the forecast. Pick up similar data from the data you have collected, and subtract the actual temperature difference of the past day from the maximum temperature of the day. Let this value be my expected temperature.
4-2 Results and consideration of [Research 2]
The forecast results based on regularity are as shown in [Fig. 3]. The difference between the forecast and the actual value was smaller in my forecast than in the Japan Meteorological Agency. In other words, I had better forecast accuracy.
However, there were days when it was unpredictable due to lack of experience, such as suddenly becoming a winter-type pressure distribution. In addition, the approach of a low pressure is not as experienced as that of a high pressure, furthermore, the weather and temperature change greatly depending on where the low pressure approaches, so I felt that it was necessary to gain more experience.
5 My proposal
The Japan Meteorological Agency relies too much on computers and thinks that detailed forecasts are not made very well. It is better to let a supercomputer do rough phenomena such as weather charts, but local forecasts such as temperature use the method based on human experience may be more accurate. It would be good if this could accurately predict the weather that would lead to a major disaster.
6 Future outlook (towards the second year)
Surprisingly, this time the forecast accuracy was better for me than for JMA. However, many issues remained.
(1)There are times when the forecast of the weather (sunny or cloudy) is not correct
(2)Inexperienced and unable to deal with phenomena with a low probability of appearance
(3)The temperature difference varies depending on the season
(4)The only data used are weather, wind speed, and normal weather maps.
From such issues, I felt that it was important to continue to collect data and gain experience. In addition, we would like to discover the correspondence with the weather and temperature by using the upper weather map, etc., and use it as a new forecast material. In the second year, I would like to develop research not only on temperature but also on weather forecasts.
7 Quote
Japan Meteorological Agency (https://www.jma.go.jp)
The changing weather is very interesting. I observe the sky every day and make simple weather forecasts. Among them, I noticed that the temperature forecast announced by the Nagano Local Meteorological Observatory (hereinafter: Japan Meteorological Agency: JMA) is different from the actual temperature. Is it possible to predict more accurately?
2 Current status (analysis of previous research)
(1) About the JMA forecast
JMA uses a supercomputer to make "numerical weather prediction". It calculates various factors such as weather and temperature. The calculation results are revised and announced by local staff.
(2) Accuracy of JMA forecast
In July 2021, as a result of investigating the expected temperature and the actual temperature of the Japan Meteorological Agency in Nagano, there was a difference of 1 ° C or more in about one-third.
(3) Research content
In order to forecast the minimum and maximum temperatures more accurately, I decided to make a temperature forecast myself. This time, I will use the data I collected to forecast the temperature and study whether the forecast accuracy will improve. We will also compare it with the Japan Meteorological Agency.
3-1 [Research 1] method
I recorded The weather, wind speed, and temperature from AMeDAS data every day, and investigated the tendency of temperature changes due to atmospheric phenomena. Enter data into a spreadsheet and categorize it to find regularity. Also, organize the weather charts into slides and save them.
3-2 Results of [Research 1]
The difference from the highest temperature to the lowest temperature on the previous day had regularity as shown in [* Fig. 1]. More than half of the atmospheric phenomena of each item are within 1 ° C around a certain value. It is a little wider on cloudy days. There was certainly regularity due to the wind speed [Fig. 2], and the change in temperature was smaller when the wind was stronger. It was found that the width in the graph depends on the wind speed.
4-1 [Research 2] method
Apply the trends and regularities obtained in [Research 1] to the forecast. Pick up similar data from the data you have collected, and subtract the actual temperature difference of the past day from the maximum temperature of the day. Let this value be my expected temperature.
4-2 Results and consideration of [Research 2]
The forecast results based on regularity are as shown in [Fig. 3]. The difference between the forecast and the actual value was smaller in my forecast than in the Japan Meteorological Agency. In other words, I had better forecast accuracy.
However, there were days when it was unpredictable due to lack of experience, such as suddenly becoming a winter-type pressure distribution. In addition, the approach of a low pressure is not as experienced as that of a high pressure, furthermore, the weather and temperature change greatly depending on where the low pressure approaches, so I felt that it was necessary to gain more experience.
5 My proposal
The Japan Meteorological Agency relies too much on computers and thinks that detailed forecasts are not made very well. It is better to let a supercomputer do rough phenomena such as weather charts, but local forecasts such as temperature use the method based on human experience may be more accurate. It would be good if this could accurately predict the weather that would lead to a major disaster.
6 Future outlook (towards the second year)
Surprisingly, this time the forecast accuracy was better for me than for JMA. However, many issues remained.
(1)There are times when the forecast of the weather (sunny or cloudy) is not correct
(2)Inexperienced and unable to deal with phenomena with a low probability of appearance
(3)The temperature difference varies depending on the season
(4)The only data used are weather, wind speed, and normal weather maps.
From such issues, I felt that it was important to continue to collect data and gain experience. In addition, we would like to discover the correspondence with the weather and temperature by using the upper weather map, etc., and use it as a new forecast material. In the second year, I would like to develop research not only on temperature but also on weather forecasts.
7 Quote
Japan Meteorological Agency (https://www.jma.go.jp)