About

BIO

Dr Michael Chase is a recently retired engineer/scientist with an academic background in theoretical elementary particle physics, and industrial experience in signal processing algorithm development.

SCOPE

This website is about “do-it-yourself” climate reconstruction from raw data, and validation of adjusted data, currently with methods for rainfall and surface air temperatures. Basic analysis can be done with a spreadsheet, sometimes adequate for rainfall, but temperature reconstructions generally require some elementary coding within a scientific programming environment such as MATLAB or R.

RAINFALL

Rainfall records often suffer from periods of missing data, and/or from inhomogeneities due to changes in location, technology and local environment. Visual comparison with data from neighbouring stations is used to identify and resolve these issues, allowing composite rainfall histories to be constructed. The intention is to place recent rainfall totals into the longest possible historical perspective.

TEMPERATURE RECONSTRUCTION

This website describes via the pages above, and illustrates via blog posts, a computer-aided iterative procedure, based on the “First Difference” method, for reconstructing the surface air temperature history of a region from its weather station data. Inter-annual temperature differences are averaged across stations, separately for each month, excluding periods deemed to be distorted by non-climatic effects such as moves and equipment/environmental changes. The resulting average temperature differences are integrated forwards and backwards in time from an arbitrary reference year, giving regional average histories of monthly average temperature variations.

TEMPERATURE VALIDATION

This website gives examples, as blog posts, of validation tests of adjusted (homogenised) temperature datasets. The tests rely on the subtraction of a suitable “reference series” to increase the signal (steps/trends) to noise (weather) ratio. Homogenised data are deemed to be “not-invalid” if they pass all of the following checks:

• Self consistency with near neighbours
• Absence of inhomogeneities
• Consistency of the adjustments with non-climatic changes in raw data

 

End of Page