SFO APPROACH FOG STUDY Noel Keene ZOA CWSU Fremont, CA
SFO APPROACH FOG STUDY Reason for study: – Fog causes significant air traffic impacts at San Francisco International Airport(SFO) – Fog is difficult to forecast Overforecast by models CWSU often forecasts earlier clearing than actual – Scenario for fog likely follows a pattern – Increased lead-time for fog events and better accuracy for total clearing will benefit air traffic operations
Data Analyzed fall/winter data for 2007/2008 and 2008/2009 Parsed out days where 12Z 500mb height was greater than 5690DM and 12Z SFO-SAC pressure gradient > -1mb and < +1mb – All non-frontal fog events occurred in these conditions – Wanted to evaluate conditions affecting fog other than the typical light winds and high pressure
Data 67 days met initial criteria – 11 fog events – 56 mornings with no fog Fog event considered to occur when visibility <1SM affected SFO approach zone – Did not use 5/8SM because archived SMB observations do not indicate visibilities <1SM
Data Evaluated several elements – 12Z OAK 500mb height – Onshore pressure gradient (SFO-SAC) – Offshore gradients (SFO-ACV and SFO-TVL) – 10Z dewpoint at SFO approach zone (SMB) – 12Z OAK 925mb wind direction and speed – Time that visual approaches occurred at SFO Typically SCT or less coverage of clds of below 3500ft Typically visibility > 3sm – Forecaster notes taken day of event
Findings Clearing time later than expected Some correlation between fog occurrence and: – 10Z SMB Dewpoint – 12Z OAK 925mb wind direction Weaker correlation between fog occurrence and: – 12Z SFO-ACV gradient – 12Z SFO-TVL gradient
Findings Many variables contribute to fog formation over Bay Area – Therefore…correlations for single variables are low Multiple Linear Regression using both 10Z dewpoint and H9 Wind Direction from 180 yielded higher R values Graphical representation yields some insight into important parameters that contribute to fog formation
Clearing Time When fog occurred…airport arrival rate never reached optimum(45/60) before 18Z Average clearing time 21:31 Two all-day events where fog transitioned to low ceilings
Clearing Time Avg
10Z SMB Dewpoint Correlation Correlation between 10Z SMB Dewpoint and Fog Occurrence at SMB – 0.26 Dewpoint values for fog occurrence – Average………………… – Standard Deviation….2.1 Dewpoint values for mornings with no fog – Average………………… – Standard Deviation….5.0
10Z SMB Dewpoint Correlation Avg
12Z OAK 925mb Wind Direction Subtracted 180 from H9 winds and took absolute value – Larger numbers represent more northerly winds Correlation between H9 wind direction and fog occurrence at SMB – 0.20 H9 wind (FM 180) direction values for fog occurrence – Average…………………..147 – Standard Deviation….35 H9 wind (FM 180) direction values for mornings with no fog – Average…………………..124 – Standard Deviation….45
12Z OAK 925mb Wind Direction Avg
Multiple Linear Regression Low P-Value and Significance F – Indicates both variables (dwpt and wind dir) have some effect on outcome (if fog occurs or not) Still fairly low correlation (multiple R) – Indicates that using only dewpoint and wind dir to forecast fog would likely yield poor results…this is fairly obvious
Northerly Flow SFO APPROACH Northerly flow is only direction where no nearby downsloping will occur Source region of northerly flow is North Bay Valleys – Average lower minimum temperatures – SF Bay can achieve saturation more readily
Conclusions Preconditions for fog during recent fall/winter seasons at SFO approach have been: – H5 Hgt > 5690DM – SFO-SAC pressure gradient -1mb Northerly flow enhances fog formation – Negative SFO-ACV pressure gradient – Significant northerly component to wind at 925mb High dewpoint temp increases chance for fog formation – Fog did not form when 10Z dewpoint temp < 45F – SF Bay does not cool as readily as inland valleys so higher dewpoint is needed Keying in on secondary elements that aid in fog formation should provide better lead-time for fog events On fog event days…clearing at SFO approach not likely before 20Z