Cloud-Top Brightness Temperature Inconsistent Predictor of Extreme Surface Rainfall in Equatorial Africa

Convective storms are an important component of the climate of equatorial Africa west of 30°E, delivering the majority of rainfall by way of mesoscale convective systems. Here, some of the most extreme values of cloud-top brightness temperature and lightning flash rate density are observed and these are often associated with intense updrafts. While strong convective (updraft) strength is typically associated with high rainfall intensity, recent global analyses of extreme events have found that storms that produce the highest surface rainfall rates have little overlap with storms that produce the strongest updrafts. A shift in the relationship between surface rainfall and updraft strength at the extremes would disproportionately affect our ability to accurately observe rainfall in equatorial Africa because rainfall observation is reliant on satellite monitoring of cloud-top brightness temperature to estimate surface rainfall rates due to poor rain-gauge coverage.

Here we investigate the relationship between extreme surface rainfall, updraft profiles, and cloud-top brightness temperature using convection-permitting (CP) regional modeling. CP modeling is not reliant on convective proxies nor convective parameterization since convective processes are explicitly captured by the model’s governing equations. In order to ensure the regional climate is well-represented by the model, rainfall totals and large-scale environmental conditions are validated with NASA IMERG and NOAA CMORPH satellite-observed precipitation datasets, and the ERA5 atmospheric reanalysis. We find that the CP model adequately reproduces climatological rainfall totals and rainfall rate frequencies over the region and is suitable for storm evaluation.

In the CP simulation, grid-point rainfall rates in the 99th percentile correlate strongly with updraft intensity (r = 0.79) and with cloud-top brightness temperature (r = -.68) with 99% statistical significance. This relation breaks down, however, at the 99.99th percentile with neither updraft intensity (r = 0.36) nor cloud-top brightness temperature (r = -.45) correlating well to rainfall rate with 99% statistical significance. Thus, the rain rate - convection relationship breaks down for the most extreme rainfall events.

Intense rainfall is common in equatorial Africa with storms producing rainfall rates in the 99.99^th percentile occurring as frequently as 25-40 times a month in just the western portion of the domain (7°E – 17°E). Out of the simulated storm population, two storm types are identified that produce rainfall rates in the 99.99th percentile. The first is ‘continental’ storms that are characterized by strong peak updraft velocities (≥10 m/s) and a cold cloud-tops (<197 K) that occur over the continental interior, especially near high terrain. The second is ‘coastal’ storms that have weaker peak updraft velocities (~6 m/s) and warmer cloud-top temperatures (>246K) that occur over the Atlantic Ocean and low-lying coastal land between 6°S and 2°N.

Continental storms are associated with high near-surface relative humidity (>90%) and a low level of free convection (100-150hPa from the surface) prior to rainfall. These storms most often generate in the afternoon over elevated terrain when surface heating and CAPE values are at a maximum.

Coastal storms are associated with higher surface relative humidity (>95%) and a lower level of free convection (50-100 hPa from the surface) than continental storms, and most often form at night or in the early morning from the remnants of dissipating storms generated further inland that interact with the moist low-level southerly flow over the Atlantic and low coastal land. Low CAPE values and a nearly-saturated lower troposphere provide the conditions that allow for extreme rainfall (≥99.99^th percentile) to be generated in lieu of a strong updraft and strong atmospheric instability. Coastal storms producing extreme rainfall rates (99.99^th percentile) occur about half as frequently as continental storms producing extreme rainfall rates in the western part of equatorial Africa (and not at all in the eastern portion), however, rainfall from these warm coastal storms may be more likely to be misrepresented or poorly predicted through satellite observation due to their high cloud-top brightness temperature, impacting the accuracy of rainfall climatologies and of severe weather forecasting and nowcasting. Moreover, we find that the presence of warm coastal storms with rainfall in the 99.99th percentile is large enough to disrupt the high correlation between rainfall rate, brightness temperature, and updrafts.