SMAP Radiance Assimilation Over Land Improves Analysis and Prediction of Tropical Cyclone Idai
Editors: Bennett Erdman
Published July 07, 2025
Soil moisture conditions can impact tropical cyclones (TCs) while they interact with land. A very wet land surface can supply additional moisture to the storm, which can lead to an increased persistence over land (sometimes known as the ‘Brown Ocean Effect’), while in contrast, drier soils lead to faster TC decay and dissipation.
Figure 1: (top) Zonal vertical cross-section of relative vorticity (s-1) for Cyclone Idai time-averaged over the period 1800 UTC 4 March 2019 to 0000 UTC 9 March 2019 following the motion of the storm, comparing the control experiment (CNTRL_ANA) and the experiment with SMAP DA (SMAP_ANA). (bottom) Composite image of the 850-hPa relative vorticity over the same time period. X-axis and Y-Axis indicate distance from storm center in meridional and zonal directions, respectively.
Here, this effect is explored in the context of numerical weather prediction (NWP) by assimilating soil-moisture-sensitive brightness temperature observations from the NASA Soil Moisture Active Passive (SMAP) mission into the Goddard Earth Observing System (GEOS) model. Assessments are then made on the impact on the analysis and prediction of TC Idai (2019) compared to a control simulation without SMAP data assimilation (DA). It was found that SMAP DA has a pronounced beneficial impact on the analyzed TC structure over land, leading to a more intense storm with a stronger, better aligned, and better-defined vorticity column (Figure 1). Even after re-emerging over the open ocean environment, the effect of SMAP DA persists, as evidenced by stronger wind speeds and a higher degree of vorticity near the surface leading to a narrower, better-defined eye (Figure 2).
Figure 2: 800-hPa wind velocity [m/s] at 00z on Mar 9, 2019, in the control experiment (top) and the experiment with SMAP DA (bottom). After an extended 5-day loop over land, the experiment with SMAP DA shows higher wind speeds and vorticity and a tighter, better-defined eye.
It was also found that assimilating SMAP observations reduces the analyzed TC size, as measured by the wind speed radius – even when the storm is not over land – which aligns better with independent observations of TC Idai. An examination of 5-day forecasts initialized from the analyses with and without SMAP DA show that constraining the initial conditions with SMAP observations leads to reduced forecast intensity errors and forecast along-track errors against observations (Figure 3).
Figure 3: TC Idai mean absolute (a) forecast intensity error and (b) forecast along-track error as a function of lead time for forecasts initialized from the control analysis (CNTRL_FCST, gray lines) and the analysis with SMAP DA (SMAP_FCST, black lines). Errors are computed against the observed intensity and track. Also shown are the number of forecasts contributing to the error estimates (gray bars). The average skill values are computed for all forecasts validated when Idai was at tropical storm strength or greater during 9 - 15 March 2024.
Beyond an assessment of the impact of SMAP DA on the TC Idai analysis and prediction, this study also includes an investigation into the mechanisms leading to the model skill improvements. This was achieved through an examination of the land surface states and fluxes in both experiments, combined with a back-trajectory analysis, that captures the interaction of the storm’s circulation with the land surface. For the case of TC Idai, it was found that through SMAP DA, the model better captures the wet surface conditions in the study region resulting from previous rain events. The wetter soil moisture conditions lead to an increased latent heat flux in the experiment with SMAP DA, which in turn results in an increased total column moisture in the storm compared to the control experiment.
The results from this proof-of-concept study provide a promising demonstration of the potential to use SMAP DA to better constrain the behavior of land-interacting TCs in the GMAO’s GEOS through an improved representation of the physical environment. This is a first step towards providing improved forecasts of these extreme weather events, which is crucial for the ability to prepare for and mitigate their socio-economic impact.
References:
Kolassa, J., Ganeshan, M., McGrath-Spangler, E., Reale, O., Reichle, R., and Zhang, S.Q., (2025), "Impact of assimilating SMAP observations over land on tropical cyclone representation: A case study of Tropical Cyclone Idai", Quarterly Journal of the Royal Meteorological Society, https://doi.org/10.1002/qj.5018
