Precipitation models

Question 1

Objectives:

Answer 1

 Forecast: real-time, flood warning

 Simulation: input for rainfall-runoff models (planning and design, impact analysis); extension of short rainfall series or generation of series for ungauged basins by regionalization

Question 2

Deterministic models:

Answer 2

 Physically based weather prediction models

 Focus on large scale weather, rainfall quite uncertain

 Forecast: NWP - Numerical Weather Prediction Model

 Simulation: GCM/ RCM – Global/ Regional climate models

Question 3

Stochastic models:

a randomly determined process.

Answer 3

 Based on statistical laws

 Forecast: e.g. MOS – Model Output Statistics (correction of RCM)

 Simulation: many methods, keeping most relevant statistical properties of rainfall process according to wanted application

Question 4

Special stochastic properties of rainfall process:

Answer 4

 Intermittent character → rainy period + dry period

 Space – time structure → spatial and temporal correlation

Question 5

Classification of stochastic models:

Answer 5

General:
1. Single site → for one station
2. Multisite → for several stations
3. Multidimensional → space-time
4. Occurrence only; amount only;
occurrence and amount

After method:
I. Alternating-Renewal-Modelle
II. Time series models
a) Markov – chains (MC)
b) Auto regression – Moving – Average - Models (ARMA)
III. Point process models
IV. Disaggregation models
V. Resampling models