The key concept of a TMY dataset is the reduction of several years of data to one representative year. In the Solargis method of preparing TMY data, we select 12 most representative months to create a typical year. The choice of representative months is made with the following objectives:
The annual and monthly sums of the TMY data should be similar to the long-term averages calculated from multi-year time series data.
The distribution of values should be preserved
Consistency of all parameters in data file must be maintained
Adequate weighting must be given to most relevant parameter when deciding on the similarity of statistical characteristics. For example, when preparing TMY data for PV applications - maximum focus is to preserve monthly average and distribution of Global Horizontal Irradiance values. Whereas when preparing TMY data for CSP applications, the maximum focus is on preserving monthly average and distribution of Direct Normal Irradiance values.