What is the Stochastic Climate Simulator?
The Stochastic Climate Simulator is a web application for hydrological and environmental modeling that:
- Generates CLIGEN Parameters: Calculates the 12-month statistical weather parameters (
PWW,PWD,Mean_P,SD_P,Skew_P,Mean_Tmax,SD_Tmax,Mean_Tmin,SD_Tmin) from historical gauge records (NOAA GHCN-Daily or custom CSV/Excel uploads). - Provides GoldSim Integration & Web Visualization: Offers interactive in-browser stochastic weather simulation and parity verification alongside a standalone 64-bit Windows DLL (
CligenWeather.dll, with source code and working examples on GitHub) that outputs synthetic daily precipitation and temperature time series directly inside GoldSim Monte Carlo models.
Purpose of Stochastic Weather Generation
Relying on a single historical weather record limits a model to the specific sequence of events observed during that period. A 30-year gauge record may not include:
- Multi-year droughts of a given duration.
- Multiple large storm events occurring within the same season.
- Warm spring periods coinciding with high antecedent moisture.
Playing back a historical record in chronological order only tests system performance against that single historical sequence. It does not quantify the probability of exceeding storage or capacity thresholds across alternative weather sequences.
Stochastic weather generation addresses this by calculating monthly statistical parameters (seasonality, wet/dry transition probabilities, rainfall depth distributions, and temperature ranges) from the historical record and using those parameters to generate multiple synthetic weather realizations.
Core Functions
The application provides three main functions:
graph TD
A["1. Historical Ingestion & Analysis"] -->|"Screen & Filter"| B["2. CLIGEN Parameter Calculation"]
B -->|"In-Browser Synthesis"| C["3A. Web Simulation & Parity Testing"]
B -->|"Copy 2D Matrices & DLL"| D["3B. GoldSim Monte Carlo Integration"]
1. Historical Ingestion & Analysis
Load historical records from global weather stations (NOAA Global Historical Climatology Network) or upload custom daily precipitation and temperature records (CSV or Excel). Review monthly averages, frost/freeze counts, diurnal temperature ranges, and record completeness before generating parameters.
2. USDA CLIGEN Parameter Calculation
The application calculates the monthly statistical parameters used by CLIGEN (the USDA Agricultural Research Service stochastic weather generator). Rather than assuming rainfall follows a normal distribution, CLIGEN models:
- Markov Chain Rain Occurrence: Conditional probabilities of a wet day depending on whether the previous day was wet or dry.
- Skewed Rainfall Depth (Pearson Type III): A three-parameter skewed distribution for wet-day precipitation depth.
- Conditioned Temperature: Daily temperature distributions conditioned on wet and dry days.
3. Execution Options: Web Application & GoldSim DLL
- In-Browser Simulator: Run daily or sub-daily simulations in the browser and inspect hyetographs, temperature envelopes, and multi-realization percentile bands (P10, P50, P90).
- GoldSim Integration (
CligenWeather.dll): Use the compiled 64-bit Windows dynamic link library with GoldSim Data elements to generate synthetic weather sequences during GoldSim Monte Carlo simulations without external text files.
Figure: The in-browser simulation dashboard showing a 10-realization Monte Carlo ensemble, statistical distributions, peak intensities, and temperature percentiles.
Figure: Simulation results showing cumulative precipitation (P10, P50 median, and P90 bounds) alongside seasonal temperature envelopes.