Best practices for precipitation sample storage for offline studies of ice nucleation in marine and coastal environments

Abstract

Ice-nucleating particles (INPs) are efficiently removed from clouds through precipitation, a convenience of nature for the study of these very rare particles that influence multiple climate-relevant cloud properties including ice crystal concentrations, size distributions and phase-partitioning processes. INPs suspended in precipitation can be used to estimate in-cloud INP concentrations and to infer their original composition. Offline droplet assays are commonly used to measure INP concentrations in precipitation samples. Heat and filtration treatments are also used to probe INP composition and size ranges. Many previous studies report storing samples prior to INP analyses, but little is known about the effects of storage on INP concentration or their sensitivity to treatments. Here, through a study of 15 precipitation samples collected at a coastal location in La Jolla, CA, USA, we found INP concentration changes up to > 1 order of magnitude caused by storage to concentrations of INPs with warm to moderate freezing temperatures (−7 to −19 ∘C). We compared four conditions: (1) storage at room temperature (+21–23 ∘C), (2) storage at +4 ∘C, (3) storage at −20 ∘C and (4) flash-freezing samples with liquid nitrogen prior to storage at −20 ∘C. Results demonstrate that storage can lead to both enhancements and losses of greater than 1 order of magnitude, with non-heat-labile INPs being generally less sensitive to storage regime, but significant losses of INPs smaller than 0.45 µm in all tested storage protocols. Correlations between total storage time (1–166 d) and changes in INP concentrations were weak across sampling protocols, with the exception of INPs with freezing temperatures ≥ −9 ∘C in samples stored at room temperature. We provide the following recommendations for preservation of precipitation samples from coastal or marine environments intended for INP analysis: that samples be stored at −20 ∘C to minimize storage artifacts, that changes due to storage are likely an additional uncertainty in INP concentrations, and that filtration treatments be applied only to fresh samples. At the freezing temperature −11 ∘C, average INP concentration losses of 51 %, 74 %, 16 % and 41 % were observed for untreated samples stored using the room temperature, +4, −20 ∘C, and flash-frozen protocols, respectively. Finally, the estimated uncertainties associated with the four storage protocols are provided for untreated, heat-treated and filtered samples for INPs between −9 and −17 ∘C.