Marine heatwaves (MHWs) are warm sea surface temperature (SST) anomalies with substantial ecological and economic consequences. Observations of MHWs are based on relatively short instrumental records, which limit the ability to forecast these events on decadal and longer timescales. Paleoclimate reconstructions can extend the observational record and help to evaluate model performance under near future conditions, but paleo-MHW reconstructions have received little attention, primarily because marine sediments lack the temporal resolution to record short-lived events. Individual foraminifera analysis (IFA) of paleotemperature proxies presents an intriguing opportunity to reconstruct past MHW variability if strong relationships exist between SST distributions and MHW metrics. Here, we describe a method to test this idea by systematically evaluating relationships between MHW metrics and SST distributions that mimic IFA data using a 2000-member linear inverse model (LIM) ensemble. Our approach is adaptable and allows users to define MHWs based on multiple duration and intensity thresholds and to model seasonal biases in five different foraminifera species. It also allows uncertainty in MHW reconstructions to be calculated for a given number of IFA measurements. An example application of our method at 12 north Pacific locations suggests that the cumulative intensity of short-duration, low-intensity MHWs is the strongest target for reconstruction, but that the error on reconstructions will rely heavily on sedimentation rate and the number of foraminifera analyzed. This is evident when a robust transfer function is applied to new core-top oxygen isotope data from 37 individual Globigerina bulloides at a site with typical marine sedimentation rates. In this example application, paleo-MHW reconstructions have large uncertainties that hamper comparisons to observational data. However, additional tests demonstrate that our approach has considerable potential to reconstruct past MHW variability at high sedimentation rate sites where hundreds of foraminifera can be analyzed.
Read full abstract