Series of charts from the paper by Feng et al.
Panel a shows the composite radar reflectivity from the observation, and panels b and c show 4-hour ensemble forecasts probabilities of a radar reflectivity composite for experiments without and with the new method for representing microphysical uncertainty. Probabilities are defined as the number of ensemble members exceeding 20 dBZ divided by the ensemble size. Panels d and e show the values of the fractions skill score (FSS) in 6-hour precipitation forecasts for thresholds of 2.0 mm/hour (light rain) and 5.0 mm/hour (moderate rain) for two experiments without (red line) and with (blue line) the new method. The FSS values are scale-dependent, and higher FSS values mean more accurate forecast. Here, the FSS values at a scale of 70 km are shown. Credit: Feng et al. [2021], panels from Figure 13 and 15
Editors’ Highlights are summaries of recent papers by AGU’s journal editors.
Source: Journal of Advances in Modeling Earth Systems

To obtain better initial conditions for numerical weather prediction (NWP) models for predicting thunderstorms, storm-scale data assimilation (SDA) is performed, which merges the information from high-resolution model runs and observations at high frequencies. However, one of the main difficulties in performing SDA is the correct representation of uncertainty in clouds and precipitation of NWP models.

The initial condition that merges the information from high-resolution model runs and observations in high frequency has to be computed fast, so that forecasts can be timely started and computed. Although more advanced microphysical schemes for simulating cloud and precipitation exist, they are rarely used in operational models due to their computational demands.  

Feng et al. [2021] calculate the differences between two model runs using different microphysical schemes and stored them in a database. Then, the obtained samples are incorporated into the SDA. To evaluate this method’s performance, SDA experiments were carried out over a one-week period in summer over Germany. The results show that the new method reduces biases in relative humidity and temperature and provides considerable improvement of short-term forecasts of radar reflectivity and hourly precipitation.

Citation: Feng, Y., Janjić, T., Zeng, Y., Seifert, A., & Min, J. [2021]. Representing microphysical uncertainty in convective-scale data assimilation using additive noise. Journal of Advances in Modeling Earth Systems, 13, e2021MS002606.

—Jiwen Fan, Editor, Journal of Advances in Modeling Earth Systems

Text © 2021. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.