Seasonal Forecasting: Accuracy Indices


Sector: Utilities

Index of power demand for cooling

TABLE OF WEIGHTING FACTORS
May/Oct May/Oct June/Sept June/Sept July/Aug July/Aug
Day Nigh tDay Night Day Night
Temp 3 1 6 3 10 5
RH 3 3 7 7 10 10
Precip 2 1 5 1 9 1

INDEX = SUM (weights)(errors) 
      =    3 T(M/O, d)  + 1 T(M/O, n) 
        +  6 T(Jn/S, d) + 3 T(Jn/S, n)  
        + 10 T(Jl/A, d) + 5 T(Jl/A, n) 
        +  3 RH(M/O)    + 7 RH(Jn/S)   + 10 RH(Jl/A) 
        -  2 P(M/O, d)  - 5 P(Ju/S, d) - 9 P(Jl/A, d) - 1 P(n)

LOGIC:

Variables Chosen

  1. Precipitation makes it feel cooler (so subtract)
  2. Temperature and RH makes it feel uncomfortable (so add)
  3. RH is more uncomfortable
  4. Set scale from 0 to 10
Months Chosen
  1. Peak power usage is more likely to break tolerance in the heart of summer so values are higher
  2. Air conditioning is usually used any time from May to October
Daily variations (day/night)
  1. RH doesn't matter if day or night
  2. Temperature does matter on day or night
  3. Precipitation kind of matters

Sector: Petroleum Industry

Sector: Fisheries


Go to Mteor 455/555 home page