Response of low-level clouds to the Kuroshio Extension front in the early summer: Field measurements

The atmospheric responses to the mid-latitude SST fronts are recognizable particularly in the cold season. In summer, the sea surface temperature (SST) gradient becomes smaller due to strong heating of the ocean surface layer, and both turbulent heat fluxes and their cross-frontal contrasts weaken. However, previous studies indicated the Kuroshio Extension (KE) can definitely affect the overlying atmosphere also in early summer. To investigate the effects of the KE front further, with a focus on low-level clouds and downward longwave radiation, an intensive atmospheric observation campaign was conducted in the Baiu/Meiyu season of 2012. Three of research vessels aligned with a latitudinal separation of 30' or 45' along 143°E across the KE front for simultaneous observations during 1-6 July. Each of the vessels moved back and forth meridionally within a half-degree section along 143°E, launching GPS radiosondes every two hours.

In the period of the simultaneous observations, the SST difference across the front was about 5 K and the SST front shifted northward by about 50 km within three days, which was not well reproduced in objectively analyzed SST datasets. Our intensive observations have revealed fine structures of the marine atmospheric boundary layer (MABL) modulated across the front. In the first half of the observation period, cool northerly airflow across the SST front rendered the MABL stratification strongly unstable over the warm KE. In the second half of the period, warm southerly airflow advanced across the front over to the cooler water, rendering the MABL stratification more stable or less unstable and contributing to the SST warming. Surface turbulent heat fluxes showed clear meridional contrasts across the SST front. While the sensible heat flux maximized near the front, latent heat flux tended to peak around 50 km south of the front where near-surface specific humidity was locally reduced probably due to enhanced vertical mixing and entrainment of the dry free-tropospheric air into the MABL. Correspondingly, a water vapor front was located 100 km south from the SST front, and between this front and the SST front cloud base was barely detected below 120 m height throughout our observation period.

The observed cloud base height in the MABL exhibited a clear tendency to be higher over the warmer water and lower over cooler water across the SST front, arising from the cross-frontal difference in the strength of vertical mixing in the MABL. This cross-frontal contrast in low-level clouds became particularly distinct under the northerlies, which is consistent with a previous study. The mode value of the downward longwave radiation (DLW) flux observed at the surface was greater by approximately 20 W/m² over warmer water than over cooler water across the SST front, which was due to the overall increase in water vapor toward the south. This effect can contribute to the maintenance of the frontal SST gradient. The difference in DLR across the front was less clear as the mean of the northerly cases since low DLR values were observed more often to the south.

Numerical simulations were performed with two kinds of high-resolution regional atmospheric model, to examine the sensitivity of low-level cloud formation to the presence of the frontal SST gradient. Those two models are found to simulate the elevated cloud deck in the MABL on the warmer side of the SST front than on the cooler side. The simulated meridional elevation is, however, underestimated relative to its observational counterpart, which is attributable to overactive vertical mixing in the model MABL to the north of the front. The substantial reduction in the cross-frontal elevation of cloud-deck height simulated under the artificially smoothed SST highlights the critical importance of the frontal SST gradient in forcing the particular structure observed in the MABL cloud-deck. The artificial SST smoothing is found to yield difference in DLR, which is 10 W/m² at most. We argue that this model-simulated effect of the SST front may be seriously underestimated because the cross-frontal SST gradient given as the model lower-boundary condition is also underestimated relative to its counterpart in our in-situ observations.