SMAP Observations Can Improve Near-Surface Humidity and Temperature in GEOS Weather Analysis

Authors: Rolf Reichle, Sara Zhang, Qing Liu, Clara Draper (NOAA), Jana Kolassa, and Ricardo Todling

Introduction

Soil moisture plays an important role in the Earth’s energy, water, and carbon cycles through its control on photosynthesis and evapotranspiration, which in turn impact atmospheric boundary layer dynamics. The accurate modeling of soil moisture is therefore critical for improving weather and seasonal climate predictions.

This study demonstrates that the assimilation of brightness temperature (Tb) observations from the NASA Soil Moisture Active Passive (SMAP) mission into the Goddard Earth Observing System (GEOS) weather analysis can improve estimates of soil moisture and, consequently, near-surface atmospheric conditions, compared to a system without SMAP assimilation.

Method

SMAP Tb observations were assimilated into a prototype weakly-coupled GEOS land-atmosphere data assimilation system during June-August 2017, along with the standard suite of atmospheric observations. For reference, a control experiment with atmospheric data assimilation but without SMAP assimilation was also conducted.

Results

The SMAP Tb assimilation improves the skill of soil moisture in the GEOS weather analysis system compared to the control experiment (Fig. 1).

Through land-atmosphere interactions, the improved soil moisture leads to better screen-level air specific humidity (q2m) and daily maximum temperature (T2mmax), with root-mean-square error (RMSE) reduced by up to 0.4 g/kg and 0.3 °K, respectively, in some regions (Fig. 2).

The improvement in the specific humidity from SMAP Tb assimilation extends into the lower troposphere (Fig. 3).

Taken together, the above results demonstrate the strong potential of SMAP Tb observations for improving global operational weather analysis and forecasting systems.

Bar graphs showing surface soil moisture skill
Figure 1: SMAP Tb assimilation improves soil moisture. Graphic shows (a) correlation (dimensionless) and (b) unbiased RMSE (m3 m-3) for surface soil moisture from (blue bars) the control experiment and (purple bars) the experiment with SMAP assimilation. Metrics are computed vs. in situ measurements from 360 sparse network sites (first group of bars) and 12 SMAP core validation sites (second group of bars) for June-August 2017. Error bars indicate 95% confidence intervals. Sparse network and core sites differ in their coverage across land surface conditions and climate zones.
global map of error in near-surface temp and humidity
Figure 2: SMAP Tb assimilation reduces the error in near-surface temperature and humidity. Maps show the difference in RMSE of simulated a) q2m and b) T2mmax with and without SMAP Tb assimilation. RMSE computed vs. in situ measurements for June-August 2017. Red colors indicate improvement from SMAP Tb assimilation.
graphs of specific humidity profile in lower troposphere
Figure 3: SMAP Tb assimilation improves specific humidity in the lower troposphere. Black dots show atmospheric profiles of the (a) mean and (b) standard-deviation of observation-minus-forecast (OmF) residuals from the control experiment for specific humidity (g kg-1) across global continental land for June-August 2017. OmF residuals were computed using radiosonde observations. Blue bars indicate the corresponding relative skill difference in OmF (a) absolute mean and (b) standard deviation in units of percent. Blue bars with negative percentage values indicate better performance with SMAP Tb assimilation compared to the control experiment.

Reference:

Reichle, R. H., S. Q. Zhang, Q. Liu, C. S. Draper, J. Kolassa, and R. Todling (2021), Assimilation of SMAP Brightness Temperature Observations in the GEOS Land-Atmosphere Data Assimilation System, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, in press, doi:10.1109/JSTARS.2021.3118595.

« GMAO Science Snapshots