search

gssti-omnibus / workshop

0 stars 0 forks source link

SARG: STOCHASTIC MODELLING OF WEATHER DATA #8

Open philusnarh opened 6 years ago

philusnarh commented 6 years ago

MARKOV CHAIN MODEL

  1. ANNUAL RAINFALL

summarize data into categories (say, low, medium, high) for each station generate Transition Probability Matrix (TPM) using Maximum Likelihood Estimator (MLE) for each.

  1. MONTHLY RAINFALL

** summarize data into WET & DRY

Schwarz criterion to choose the best optimum order Use Chapman-Kolmogorov equation to determine the ergodic distribution matrix of the system.

asarefi commented 6 years ago

Lets check on rain given and rain given dry using markov chain.

philusnarh commented 6 years ago

@asarefi just a reminder,

  1. How to calculate the annual rainfall from the daily data
  2. How to calculate the annual temperature from the daily data
  3. Same for monthly

Also,

Can we consider the rainfall in four states:

state 1 -->> = 0.0 state 2 -->> 0.1 - 3.9 state 3 -->> 4.0 - 27.9 state 4 -->> >= 28

asarefi commented 6 years ago

Rainfall We sum daily rainfall amount to generate either monthly or annual total. Temperature We calculate mean of over the period of interest either montly or annual

Rainfall can be calculated in the suggested threshold depending on what you want to achieve. Can we please discussed that further before concluding on those thresholds at your earliest convenience.