A view on recent ice-nucleating particle intercomparison studies: Why the uncertainty of the activation conditions matters

Abstract. Ice-nucleating particles (INPs) play a crucial role in cloud formation, influencing cloud phase, lifetime, and the onset of precipitation. Consequently, microphysical processes involving INPs strongly affect the radiative properties of clouds. However, when multiple INP counters measure simultaneously, notoriously high deviations between instruments in the range of 1 order of magnitude are commonly observed. These differences occur in ambient atmospheric measurements as well as in laboratory studies. A regularly overlooked reason for these discrepancies may be related to uncertainties and errors in the temperature measurement. As the activation of INPs is a strong function of the nucleation conditions, relatively small inaccuracies in the temperature measurement may lead to significant over- or underestimations of the INP concentration. In this study, we have explored this effect as a potential reason for the differences observed among INP counters participating in 10 intercomparison studies that were published within the last 10 years. The stated temperature uncertainty of instruments participating in these experiments ranged from ± 0.1 °C to ± 1.5 °C, and was most commonly specified as ± 0.5 °C. Potential deviations resulting from typical temperature errors were compared to the reported level of agreement among intercompared methods. As a measure of the potential INP error due to nucleation temperature error, we defined the error factor (EF) as the quotient of the ice nucleation activity at the actual nucleation temperature divided by the ice nucleation activity at a potentially erroneously measured temperature. Respective EFs were calculated for five distinct activation spectra based on four INP parametrizations and one compilation of atmospheric INP data. EFs were between 1.1 and 3.2 for temperature errors of ± 0.5 °C, and between less than 2 and larger than 10 for temperature errors of ± 1.5 °C. EFs calculated from parametrizations of aerosols that are highly ice nucleation active were significantly larger than those derived from atmospheric data; although the effect was found to be still as large as a factor of 10 for certain temperature ranges in atmospheric activation spectra at a temperature error of ± 2 °C. When comparing two INP instruments, measurement biases may be of opposite direction, thus resulting in expected differences of up to the product of both EFs. We found that opposite biases of +0.5 °C and −0.5 °C can therefore typically explain differences of a factor of 2, while opposite biases of +1 °C and −1 °C can theoretically explain differences of factors up to 5 or even 10, which is in the order of discrepancies typically reported in the literature on INP intercomparisons. These results highlight the need to carefully assess and report on uncertainties of the ice nucleation activation conditions.