Abstract
系集預報的精神,是希望藉由這些多個單位模式的重複預報,消化篩選出目標資訊。藉由多個不同的系集成員預報,可以彌補單一模式預報的不足,期望能包含模式預報的不確定性,並且將不確定性量化,以提供未來的預報機率。本組之系集預報資料,來自24 個不同研究單位,各單位的超級電腦利用其各自的運算系統,演算出未來72⼩時各雨量站的預測⾬量,或是某颱⾵的預測路徑。
各模式每6⼩時即會產生一組預測區間為72小時之數據,因此,使⽤者面對的是龐大且繁雜的資料,除了需要有效的方法消化資料之外,也需要利用視覺化技術,使得重要關鍵資訊能快速呈現,以作為政府單位之決策依據。
本系統將系集預報的預測雨量與歷史淹水門檻值做比對,以投票的票數概念,針對全台各地,呈現出24 個預測模式中有幾個模式的⾬量超過淹水門檻,用以表達可能的淹水概況。
我們使用Tableau設計此次視覺化成果,總共設計了兩組畫面,第一組畫面以投票的票數概念,呈現了全台灣各縣市的可能淹水分布;第二組畫面則更詳盡地顯示了24個模組針對某單一縣市的投票內容。
各模式每6⼩時即會產生一組預測區間為72小時之數據,因此,使⽤者面對的是龐大且繁雜的資料,除了需要有效的方法消化資料之外,也需要利用視覺化技術,使得重要關鍵資訊能快速呈現,以作為政府單位之決策依據。
本系統將系集預報的預測雨量與歷史淹水門檻值做比對,以投票的票數概念,針對全台各地,呈現出24 個預測模式中有幾個模式的⾬量超過淹水門檻,用以表達可能的淹水概況。
我們使用Tableau設計此次視覺化成果,總共設計了兩組畫面,第一組畫面以投票的票數概念,呈現了全台灣各縣市的可能淹水分布;第二組畫面則更詳盡地顯示了24個模組針對某單一縣市的投票內容。
Ensemble forecasting is a numerical prediction method that is used to attempt to generate a representative sample of the possible future states of a dynamical system. Sometimes the ensemble of forecasts may use different forecast models for different members, or different formulations of a forecast model. Thus, the predicted ensemble spread can be obtained and the amount of spread should be related to the uncertainty (error) of the forecast.
In this project, we use 24 different forecast models. Each model will predict the estimated rainfall of every rainfall stations for the next 72 hours, or the predicted path of a typhoon. And each model will produce one forecast every 6 hours. Users are facing a great amount of data. Thus data visualization can be a important solution to reveal the critical information, and to serve as the basis for decision making for government.
In this project, we compare the estimated rainfall with the threshold of historical flooding. With this concept of voting, we can get the votes of how many estimated-value are over the threshold within 24 models and thereby realize the future flooding situation of Taiwan.
Tableau Software is used to design and complete our visualization. We’ve designed two screens: one shows the general flooding situation of cities and counties in Taiwan on the basis of votes; the other one shows the voting detail of the 24 models for each city (county).
In this project, we use 24 different forecast models. Each model will predict the estimated rainfall of every rainfall stations for the next 72 hours, or the predicted path of a typhoon. And each model will produce one forecast every 6 hours. Users are facing a great amount of data. Thus data visualization can be a important solution to reveal the critical information, and to serve as the basis for decision making for government.
In this project, we compare the estimated rainfall with the threshold of historical flooding. With this concept of voting, we can get the votes of how many estimated-value are over the threshold within 24 models and thereby realize the future flooding situation of Taiwan.
Tableau Software is used to design and complete our visualization. We’ve designed two screens: one shows the general flooding situation of cities and counties in Taiwan on the basis of votes; the other one shows the voting detail of the 24 models for each city (county).