Lagged linear regression may be useful as one way to explore statistical relationships and teleconnections between climate variables on different timescales and in different regions. The figures on these pages allow an interactive exploration of such relationships in observations and models, in which lagged linear regression has been investigated:

where Y(t) is the dependent varable at time t, X is the independent variable with lag dt, a is the regression coefficient, with some noise, n. By selecting from the options below teleconnections between different climate variables and regions can be examined through each of the figures presented. In each case figure 1 shows the correlation between the regionally averaged dependent variable (for example global mean surface air temperature) and the regression model using the independent variable (for example sea surface temperature in each location over the globe) with a lag (from zero to seasons or years) as a predictor. Figure 2 shows the regression coefficient at each location and figure 3 shows the time series of the dependent variable. The stippling gives an indication of regions in which teleconnections are not statistically significant (they are locations where the 1-sigma uncertainty on the regression parameter is as large or larger than the magnitude of the regression parameter itself).