[M-GI29] Data assimilation: A fundamental approach in geosciences

[MGI29-P04] Toward assimilation of high-precision 3-D lightning location data for severe thunderstorm forecast

*Yasumitsu Maejima¹, Futo Tomizawa^2,1, Tomoo Ushio³, Takemasa Miyoshi^1,4,5 (1.RIKEN Center for Computational Science, 2.Graduate School of Engineering, the University of Tokyo, 3.Graduate School of Engineering, Osaka University, 4.University of Maryland, College Park, 5.Japan Agency for Marine-Earth Science and Technology)

Keywords:Data assimilation, Numerical weather prediction

In recent years, a high-precision lightning observation system named “BOLT” was developed (e.g. Yoshida et al. 2014), observing every 10^-6 seconds 3-dimensional lightning locations at 100-m resolution. The original data have a 70-km range in the horizontal and a 20-km range in the vertical. To assimilate the lightning location data with a regional numerical weather prediction model “SCALE” (Nishizawa et al. 2015), an observation operator to convert the model variables to lightning data is needed. In general, lightning is triggered by large electric charge originated from graupel collisions with strong updraft. Therefore, we investigated a relationship between the lightning observation data and the model graupel data. First, we performed a 30-seceond-update, 100-m-mesh SCALE-LETKF (Miyoshi et al. 2016, Lien et al. 2017) experiment with every 30-second Phased Array Weather Radar at Osaka University (Ushio et al. 2014), and created every 30-second model data. For the lightning observation data, we calculated every 30-second lightning frequency given by counts per second per model grid. We investigated a scatter plot of the lightning frequency and the model graupel mixing ratio.

Using all lightning and graupel data, we could not find a clear relationship. However, we found a high correlation between the lightning frequency and the graupel mixing ratio by applying the following criterions on lightning generation described in Takahashi (1978): w > 1, 10 > qc > 0.01 and -3 > T > -30, where w, qc and T represent vertical motion [m s^-1], cloud water content [g m^-3] and temperature [℃], respectively. The result may help develop an observation operator to improve severe thunderstorm forecast.