Abstract:
본 발명의 대기 분석 장치는, 대기를 관측한 실측 자료를 입수하고 실측 자료를 분석하여 실측 유니트 자료를 출력하는 실측 유니트와, 대기의 상태를 모의하여 모의 유니트 자료를 출력하는 모의 유니트와, 실측 유니트 자료와 모의 유니트 자료를 비교 연산하고 대기의 분석 자료를 산출하는 산출 유니트를 포함한다.
Abstract:
The present invention relates to a method for optimally improving a surface water simulation for physics options by using a micro genetic algorithm of a genetic algorithm in a Noah multi-physics options model. The present invention not only improves the accuracy of the model but also maximizes the accuracy of the model in a wide target area for water variables with large deviation according to the difference of a coating state and surface soil like evapotranspiration and runoff.
Abstract:
본 발명은 Noah 다중 물리 지표모델에 유전 알고리즘 중 마이크로 유전 알고리즘을 이용하여 지표 물리 옵션에 대한 최적의 지표 물수지 시뮬레이션을 개선하는 방법에 관한 것으로, 모델의 정확도를 개선할 수 있을 뿐만 아니라, 증발산이나 유출량 같은 지표의 토양 및 피복상태의 차이에 따라 지역적 편차가 큰 물수지 변수들의에 대하여 광범위한 대상지역에서도 모델 정확도를 극대화 할 수 있다.
Abstract:
According to the present invention, an annual vegetation predicting method includes: a step of separately analyzing a correlation between a plurality of meteorological factors and the main ingredients of a plant activity phase mode (PM) and a plant activity intensity mode (AM) which are disassembled from the past vegetation index through an empirical orthogonal function analyzing method; a step of separately selecting at least one predictive meteorological factor determined to have a statistically meaningful correlation among the meteorological factors based on the correlative analysis of each prediction of plant activity intensity and plant activity phase; and a step of separately predicting the plant activity intensity and the plant activity phase based on the past data relating to the plant activity intensity and the plant activity phase, and the present and past data for the selected prediction meteorological factor.
Abstract:
The present invention relates to a method for obtaining a best independent model and improving simulation accuracy which maximizes the simulation accuracy for a single or a plurality of target variables in an available physics parameterization technique by applying a model optimization algorithm to a multi-physics model. The present invention can use a minimized computing system resource capable of maximizing the simulation accuracy of the single or the plurality of target variables among the physical parameterization schemes.
Abstract:
PURPOSE: A genetic algorithm interface system applicable for a numerical model is provided to improve the accuracy of rainfall forecast by being applied for a weather forecasting model. CONSTITUTION: A system for forecasting rainfall includes a data inputting part(110), a model inputting part(120), a first parameter generating part(130), an ETS storing part(140), a second parameter generating part(150), and a convergence analyzing part(160). Rainfall data is input into the data inputting part. A forecasting model based on convective parameterization is inputted into the model inputting part. The first parameter generating part generates arbitrary parameters based on the input rainfall data. The ETS storing part stores rainfall forecast by implementing the forecasting model based on the parameters. The second parameter generating part evaluates the stored ETS based on a genetic algorithm and generates enhanced second parameters. The convergence analyzing part implements the forecasting model based on the second parameters and analyzes the convergence of parameter values.