Hele tekst
Weekly Meeting 18Aug2017
Forecast SW (LSR3) with the linear regression model o Using 5o×5o box-averaged anomaly
o Using GEFS predicted SCP as a predictor o Forecasts for MAM
o Forecast skill assessed based on cross-validation
• Anomaly correlation
• Hit rate (3 categories, 33% each)
o Comparison between the skills with 5o×5o box- averaged anomaly vs. 0.5o×0.5o grid
o Check hit rate
GERELATEERDE DOCUMENTEN
[r]
The 45-day CFSv2 hindcast data (1999–2012) were extended to 2014 by including the real-time forecasts (2013–2014).. Areas were selected for constructing
Correlation between weekly mean CFSR SCP and LSR o Total field. o
from the CFSR
Percentage Percentage PercentagePercentage
o LSR: Local Storm Report (hail, tornado, damaging wind) o Anomaly correlation (AC) for 7-day mean.. o 3-month shift windows o Conditional AC (SCP
Weekly Meeting 07Jul2017.. Co-variability between SCP
The correlation (GEFS SCP, OBS LSR) can be increased by averaging both SCP and LSR anomalies in a larger domain prior to calculating the correlation. The skill for forecasting
