Are long-term trends in mixed layer depth influencing North Pacific marine heatwaves?
The following summarizes work published at the Bulletin of the American Meteorological Society.
A recorded summary of the major results can be found under the Presentations tab.
See also our recent Nature Comment on marine heatwave definitions
Major Results:
Long-term mixed layer depth shoaling trends in the North Pacific increased the likelihood and intensity of the 2019 Blob 2.0.
The observed trends are consistent with internal variability, but future mixed layer depth shoaling is expected in response to anthropogenic forcing.
Future changes in the mean mixed layer depth will increase the intensity and frequency of future marine heatwaves.
An approximation for the mixed layer temperature tendency when considering only local sources and sinks (i.e., neglecting advection) can be written as:
Equation 1 Mixed layer heat budget. Term I represents the impact of surface heat flux (Q) anomalies on the temperature tendency. Term II represents the impact of mixed layer depth (h) anomalies and is large in the North Pacific during boreal summer. Term III is small. Term IV quantifies the influence of entrainment from below.
Throughout the North Pacific, the mean mixed layer depth (MLD; denominator of Terms I-III) is projected to decrease in response to anthropogenic climate change. A shallower mean MLD would result in a larger temperature response for the same heat flux and/or MLD anomaly, potentially increasing the likelihood and intensity of marine heatwaves.
Since 1980, observations generally show decreasing MLDs throughout the North Pacific during boreal summer (Fig. 1c-d).
There is some spatial correspondence between observations and climate models forced with historical radiative forcing, especially with CMIP5 models. This suggests there may be a role for anthropogenic climate change in the observed MLD trends. However, the observed trends tend to be larger than the forced trends and often fall outside the model ensemble spreads. Therefore, observations likely contain a significant contribution from internal variability.
Figure 1 Summer MLD trends in observations (c-d) and climate models forced with historical+RCP8.5 radiative forcing (e-h).
Figure 2 (a) Upper: Summer MLD anomalies averaged in the red box in Fig. 1c for two sets of observational reanalyses (red/orange), CESM1-LE (black/gray), 13 CMIP5 models (blue). Shading represents full range (minimum-to-maximum) of values for each respective ensemble. Lower: MLD forcing terms calculated for CESM1-LE ensemble mean. (b) Probability distributions of CESM1-LE MLD anomalies averaged in same red box in the “present” (2005-2034; blue) and “future” (2070-2099; red) . Vertical black lines show the summer 2019 MLD anomalies in GODAS ocean reanalysis and Argo gridded data.
Looking towards the end of the 21st-Century, climate models project a robust MLD shoaling of about 4m in the Northeast Pacific (Fig. 2a; black and blue), which is primarily due to an increase in surface heat flux buoyancy forcing (Fig. 2a; dotted green).
Large negative MLD anomalies contribute to surface warming through Eq. (1) Term II, particularly during boreal summer. As a result of anthropogenic climate change, the distribution of realistic MLD anomalies are shifted towards shallower values (Fig 2b). These changes increase the likelihood of extreme negative MLD anomalies and also increase the likelihood of marine heatwaves in the future.
For example, the extreme 2019 negative MLD anomalies (black lines), which were an important component in driving the recent North Pacific marine heatwave (Amaya et al. 2020), were a very rare occurrence relative to the present climate. However, in the future, anomalies of this magnitude will be more common.
Figure 3 The contributions of Terms I and II from Eq. (1) to the mixed layer temperature as calculated in CESM1-LE, given a surface heat flux anomaly (Q’) and MLD anomaly (h’) of the same magnitude as observed summer 2019 in the Northeast Pacific (red box, Fig 1c). The CESM1-LE projected change in the mean surface heat flux and mean MLD were used in a 30-yr sliding window for Qbar and hbar, respectively.
To illustrate the MLD effect on future Northeast Pacific marine heatwaves, we calculate Terms I and II from Eq. (1) using the projected changes in summer mean surface heat flux (Qbar) and mean MLD (hbar) from CESM1-LE (Fig. 3).
By holding Q’ and h’ fixed at observed summer 2019 levels (7.8 Wm^2 and -6.2m, respectively), we quantify how much more warming these anomalies would produce when acting on a shallower future mean MLD.
By the end of the 21st-century, Term II contributes ~4.5˚C more warming than it did in summer 2019, which is due to the decreasing mean MLD and increasing mean surface heat flux.
These changes may be partially compensated for by damping from entrainment. We plan to analyze these features in greater detail in a future study.