Reanalyzed datasets can be used to create a coherent, long-term record of past weather by compensating for the inherent biases of the different instruments used to take measurements at different points in the history of weather observation. Reanalysis involves a variety of data synthesis methods that are used to incorporate different datasets into one regularly spaced grid. This ‘gridding’ process makes it easier to compare observations from different datasets, and preserves the original data collection model to ensure that the historical record isn’t influenced by artificial factors.