Fig. 5

Left panel: the emulator expectation and credible intervals as in Fig. 4, however now the observation z plus observed error has been included as the horizontal black solid and dashed lines respectively. The implausibilities I(x) are represented by the colours on the x-axis: red, yellow and green for high (I(x)>3.5), borderline (3.5<I(x)<3) and low (I(x)<3) implausibility respectively, with the green interval defining the non-implausible region \(\mathcal {X}_{2}\) for the second wave. Right panel: the second wave is performed by evaluating two runs located within \(\mathcal {X}_{2}\). The emulator becomes more accurate over \(\mathcal {X}_{2}\) and the implausibility more strict, hence defining the smaller non-implausible region \(\mathcal {X}_{3}\), given by the green interval. As the emulator is now far more accurate that the observed errors within \(\mathcal {X}_{3}\), additional runs will not significantly reduce \(\mathcal {X}_{3}\) any further, and hence the history match is complete