AIML_Laboratory

Python AIML Repo

Terrain predicted meteorological variance of precipitation.

Paul Dunn

Executive summary

Meteorological data such as precipitation and temperature can be modeled predictively based on time series methods. However, localized data will vary based on topography.

Rationale

Localized, site specific meteorological information is useful for several operational purposes such as emergency response, risk maagement, tactical defense, and agriculture among others. It is impossible to have weather stations at all locations, but there are methods for spatial predictive modeling to estimate spatiotemporal results for areas of interest (AOI)

Research Question

Spatially predict site specifc temperature and precipitation based on historical time-series data and Empirical Bayesian Kriging 3D (EBK) geostatisticasl methods.

Data Sources

Meteostat.

Methodology

Time series decomposition of ~20 years of monthly preciptation data for several sites disttibuted by climate. EBK 3D model used to interpolate predicted site samples

Results

Time series preidiction needs to account for drought in future iterations.

Next steps

Spatio-temporeal database development to show both time and elevation layers over area of interest (AOI).

Outline of project

https://github.com/1969Paulie/AIML_Laboratory/

Paul Dunn [email protected]