Figure 15. Precipitation (averaged from 5S–5N) from TRMM (left), S-SHiELD without MLO (center), and S-SHiELD with MLO (right), for initializations at (top) 15 October (middle) 8 November and (bottom) 6 December 2011.
Klingaman and DeMott (2020) found that climate models exaggerate the effect of ocean coupling on the MJO by over-intensifying the MJO in El Niño years. S-SHiELD does not have a coupled dynamical ocean and nudges towards climatology, and so can only represent the El Niño-Southern Oscillation (ENSO) state at initialization; indeed, the DYNAMO period was during a La Niña event (see https://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ONI_v5.php ). Hence, this ENSO contamination of the link between ocean coupling and the MJO is not present in S-SHiELD.
The diurnal cycle of precipitation is another challenge for climate models. Covey et al (2016) found nearly all climate models, even the 30-km resolution GFDL HiRAM, struggle with both the phase and amplitude of the diurnal cycle, especially over land and during boreal summer. Figure 16 presents the JJA diurnal cycle from a 10-year S-SHiELD simulation with MLO SSTs nudged towards climatology, with results from 13-km SHiELD hindcasts shown for reference. We find that the observed phase of the diurnal cycle is beautifully matched by S-SHiELD, over both land and ocean. Most notably the CONUS evening maximum of precipitation is reproduced. However, the amplitude of the cycle is biased low over land areas, possibly due to the inability of S-SHiELD’s 25-km grid to produce the propagating mesoscale convective systems characteristic of heavier warm-season precipitation events. This appears to be a resolution effect as 13-km SHiELD reproduces both the correct phase and amplitude of precipitation. We also find that the majority of precipitation in S-SHiELD (55% globally and 80% between 20S and 20N) is from the SAS convective scheme, although this does not adversely affect the phase of the diurnal cycle. S-SHiELD does have the correct phase and amplitude (albeit slightly too high) of the diurnal cycle of 2-m temperature over land (Supplemental Figure S3).