김동훈(Dong-Hoon Kim)'s Blog

- Quantitative analysis - MERR (Mean Error) : What is the average forecast error ? MERR (`-oo~oo`) = `1/n sum_(i=1)^n (F_i - O_i )` RMSE (Root Mean Square Error) : What is the average magnitude of the forecast errors? RMSE (`0~oo`) = $sqrt((1)/(n)sum_(i=1)^(n)(F_i-O_i)^2)$ CORR (Correlation Coefficient) : How well did the forecast values correspond to the observed values ? CORR (`-1~1`) = `( sum..

CCSM의 대기모형 결과 내에는 증발량에 관련된 QFLX(Surface water flux, kg/m^2/s) 와 강수량인 PRECC(Convective precipitation rate, m/s), PRECL(Large-scale (stable) precipitation rate, m/s) 가 있다. 우선 증발량은 다음과 같이 구한다. E(mm/day) = QFLX * 86400.0(1/sec) * 1000.0(mm) / 1000.0(kg/m^3) 강수량은 다음과 같이 구한다. P(mm/day) = (PRECC+PRECL)*86400(1/sec)*1000.0(mm) 이렇게 구한 결과에서 P - E 를 계산하면 된다.

참고: here A Running average filter The similarity between NCL and fortran-90 syntax can readily be seen in this example. Note that an NCL built-in function (runave) exists that can perform this calculation. The example is still useful however in demonstrating function construction. The following built-in NCL functions are used: dimsizes(to check the number of elements in a vector), isatt (to chec..