Tropical cyclogenesis prediction using lagged average forecast method with global nonhydrostatic atmospheric model

We performed the lagged average forecast of 30 day simulations with a horizontal resolution of 14 km to investigate feasibility of extended-range forecast of tropical cyclogenesis (TCG) and to understand what factors affect the forecast lead time of TCG. Simulations were conducted using global nonhydrostatic atmospheric model (NICAM) from middle of May to middle of June 2012 (32 initial conditions), which covers 5 and 4 TCG cases over north western and eastern pacific, respectively. During this period, the intraseasonal variation such as Madden-Julian oscillation is not so active. For the western pacific cases, the TCGs associated with westward-moving disturbance (easterly wave) were captured by the NICAM 1-2 weeks prior to the genesis. The environmental factor like genesis potential index is also reproduced for lead time of 1-2 weeks, indicating the NICAM captures the large-scale condition which is favorable for TCG. For the eastern pacific cases, the NICAM reproduces some TCG prior to 3-4 weeks prior to the genesis, but the model tends to form the multiple tropical cyclones that are artificial. This may be related to the overprediction of easterly waves.