I departed Boulder on Tuesday afternoon for Norman, Oklahoma. At that time, the plan was for a mid-evening departure and then drive toward south Florida for a possible landfall. Both mobile mesonets (MM) arrived in Norman with apparently nonfunctional data acquisition systems. A quick check revealed that the laptop in Probe3 (P3) was at fault and another was substituted. Troubleshooting Probe4 (P4) was more difficult. Eventually, it was determined that the voltage being supplied to the system was below minimum thresholds and further checks revealed the causes. Both systems were fully operational by about 2:30 am on Wednesday morning. Notes on Hurricane Georges intercept
David Blanchard
NOAA/NSSL and OU/CIMMS
An early morning departure was planned, but not executed. By the time we had all the staff together, gone over the mission requirements, packed the vehicles, and completed other mundane chores it was close to noontime. We departed shortly thereafter. Our route was south on I-35 to Dallas, then east on I-20 to Shreveport, southeast on I-49, and east on I-10. After much debate with team members and nowcasters about how far we should attempt to drive in one day, we stopped in Hammond, Louisiana, at about 2:30 am. This was our second late night in a row and was part of our decision making process. Our final instructions were to arise early and be on the road as quickly as possible so that we could make our rendezvous with Georges in Miami.
By early morning, it was becoming clear that Georges was going to continue on a more WNW direction rather than recurve into south Florida and that a Gulf of Mexico landfall was more likely. Given this new scenario, we decided that a leisurely pace was more appropriate until a better determination of where landfall might occur was made. Our first stop of the day was in Pass Christian, Mississippi, where Hurricane Camille made landfall in 1969. An "X" was drawn in the sand next to the words "Georges is coming." It was a joke, but turned out to be eerily accurate. Photographs of the MM were taken at this location since it was the first time the vehicles had made it to the Gulf of Mexico.
Late in the afternoon, we received new instructions to make haste for Naples, Florida, so that we could collect data on the eastern edge of the storm as it brushed the southwest coast of Florida. After checking mileage from our current location, the time of day, the expected time of passage of Georges, and the possibility that we would have to make a quick turnaround to get back to the northern Gulf coast, team members emphatically urged the forecasters to reconsider their decision. The nowcasters heard our arguments and were persuaded that we had made valid points.
Our new destination for the evening was Tallahassee. This location gave us access to weather data at both the Florida State University Department of Meteorology and the National Weather Service Office. The FSU location was more convenient to our motel and we selected that site for our morning briefing. Students and staff in the Department were very helpful and excited to have us visit. We received forecast data from the Krishnamurti tropical storm model which suggested recurvature into the northeast Gulf. Since landfall location was still uncertain, the decision to spend another evening in Tallahassee was made. Another stop at FSU was made during the late afternoon to get updates. Based on this new information, we decided that we would depart Tallahassee in the morning for Pensacola, Florida. Motel reservations were made for the following night in Pensacola, and in Baton Rouge, Louisiana for the following two nights. (The Baton Rouge reservations would eventually be canceled.)
A late morning departure was made from Tallahassee for Pensacola. Upon arrival in Pensacola, we briefly checked into our rooms so that we could get a hurricane update from the Weather Channel. The crew carrying the mobile 10-m wind tower continued on west to Gulfport, Mississippi for the night. The two MM teams then began data collection. P4 began a westbound transect along I-10 to Gulfport and then back to Pensacola. P3 went south to the ocean and began an eastbound transect. At this time, sea breeze convection began to rapidly develop. A funnel cloud was noted over the bay and photographed. A tornado warning was also issued for the Crestview, Florida, area. Large storm-generated surf was already present at this time and the cirrus shield from hurricane convection was clearly visible to the distant SE-S. ThetaE dropped from 361K to 345K in convection and strong, gusty winds were present in the thunderstorms. Blowing sand occasionally obscured the road but driving conditions were not impaired. A local law enforcement officer hailed us to ask what we were doing and promised us full cooperation in any way we needed it to fulfill our mission.
P3 continued along the coast until daylight began to fade then turned north along the eastern perimeter of Eglin Air Force Base. The final leg was completed doing a westbound transect along I-10 in light to moderate rain. Lightning was visible and a count was made of all strokes that we could see. Lightning frequency was about 1 per minute. We assume these were all in-cloud flashes since no ground strokes were observed.
Data collection was from 980926/1755 through 980927/0105 UTC.
At 5:15 am, we received a wakeup call from Jerry Straka ("Rise and Shine, it's showtime!"). Data collection started at 980927/1122 UTC and continued nonstop until 980928/1930 UTC. The mission of P3 was to drive west into Mississippi, then do circuits from Pascagoula to Lucedale to Mobile to Pascagoula. P4 would do similar circuits but farther to the west. During the first circuit, we began to have trouble with our winds which were related to faulty GPS readings. During a break in the light rain, we changed data acquisition systems and the problem appeared to be fixed. Unfortunately, the problem reappeared some time later. A second stop was made to replace the GPS antenna system. Again, the problem appeared to be fixed only to reappear later. Since we had replaced all the GPS-related systems, we concluded that the problem must lie elsewhere. Our best guess is that the tall trees and narrow roadways may have caused problems with GPS signals or that there was something in the local geology that was giving bad readings. In any case, once we got back onto the interstate highway, the problem went away. P4 experienced the same sort of problems.
The westbound leg of our second circuit was made along US90 on the beach front. We dropped south from I-10 into Pascagoula then went west into Gulfport. The early phases of the storm surge were already evident at this time as water was nearing the tops of many docks and piers. Surf was fairly calm since the barrier islands were quite effective at blocking the waves. We then turned north on US49 to Wiggins, Mississippi. An eastbound leg was completed on SR26/US98 into Mobile. Based on current information, it was decided that collecting data east of Mobile and Pensacola might be useful so we followed I-65 to the northeast then south on SR59 to Bay Minette, Alabama. A quick check of the Weather Channel in a local motel convinced us that we needed to return to our original location before conditions deteriorated any further.
During our first circuit we had briefly stopped at the Mississippi Travel Center, located just west of the Mississippi/Alabama state line. We pulled into the Travel Center (TC) for the second time late in the afternoon. The TC was closed with a posted sign, but was staffed by a security guard, her caretaker husband, their daughter, and off-duty security guard. We were invited in to dry off, drink coffee, and talk about the work we were doing. We decided this would be an excellent base for our operations during the next 12-24 hours and accepted their offer. The TC had well-landscaped grounds with excellent drainage and we were confident that localized flooding would not be a problem. The large landscaped area also provided good site exposure for collecting data while stopped. Since we would be doing a considerable amount of night time work in conjunction with the NOAA P-3 aircraft, it was decided to rest, eat, and nap for a few hours at the TC.
The power was lost in this area early in the evening but phone service remained up during the storm. We maintained contact with our nowcaster and began a series of east-west transects late in the evening. Our strategy was to keep one MM stationary and move the other MM a short distance (0.1 mile increments) every two minutes. This would give us data with varying baselines so that we could measure phenomena on a variety of scales. The NOAA P-3 aircraft was airborne and collecting data in this same location. After completing a set of westbound and then eastbound transects, we returned to the TC at around 2:30 am. During these transects, both crews observed that the recorded 6-s winds seemed to be below our best visual estimates and we decided that the vector averaging may have underestimated the actual wind speeds. On the other hand, we had noted earlier in the day that there appeared to be a significant amount of speed shear between 1 m, 3 m, and tree top winds, so it is not entirely clear if the recorded winds are underestimates or not. (Post-storm checks of the data indicates only small differences between the vector and scaler averages of the wind. Additional hardware checks indicated that there was no error in the data collection system.)
As we finished the transects, the outer edges of the eyewall were coming ashore and visibility along the highway was deteriorating. We returned to the safety of our base for the duration of the dark hours. One MM crew slept in their vehicle; the other crew took short naps in the TC. Around 4 am, water started leaking through the roof of the TC since many shingles had been ripped from the exterior of the building. All personnel in the TC were moved from the south-facing rooms to the safer and drier north-facing rooms. Also around this time, a television crew from Mobile, Alabama (NBC, Channel 15) drove through and spotted the MM. They were interested in doing an interview and I consented. I was able to get across the goals of the overall mission, the large number of collaborative organizations, and the large array of instrumentation present. They were happy with the interview. I moved from the TC out to the MM since the data system had been running on battery power for about 3 hours and I wanted to start the engine to ensure that we did not drain the batteries.
I stayed with the car until first light, then awakened the other crew members. By 1150 UTC, we were on the road. Our goal was to do transects along and through the eyewall as long as wind speeds remained below our safety thresholds. Since the eye had now been ashore for almost two hours, we encountered winds of only 30-40 knots with some higher gusts. A wind shift was approximately collocated with the pressure minimum (964 mb) and winds went from southeasterly to northerly over a short distance. We continued the transect to mile marker 24 (northwest of Pass Christian), then reversed direction for an eastbound transect. On this pass, we noted two pressure minimums and the same wind shift. During these transects, the majority of observed damage was confined to large billboards that had been shredded by the wind or had their posts snapped. Mobile homes along I-10 showed little evidence of damage. Pine trees had shed large amounts of needles onto the highway; this was the only debris we encountered. Construction barrels (those large, sand-filled orange objects) were, by and large, still standing although a few had been pushed into adjacent lanes. Since there was very little traffic on the road, this presented no hazard.
At mile marker 46 (D'Iberville), we went south on SR609 to Ocean Springs. The road had only minor flooding at one intersection (no waves or moving water) and passage was considered safe. We turned west on US90 through Ocean Springs, then crossed the causeway over the bay. After exiting from the west end of the bridge, we encountered deeper water with waves and chose to abort the transect. At this time, another vehicle came up behind us in the adjacent lane and decided that they could drive through the water. Wrong! Within a few tens of meters, they had drowned their engine. A television crew at this same location hopped into the ~1/2 meter deep water and pushed the car back onto dry ground. This deep water represents the only serious hazard we encountered and we found it to be easily avoided. We retreated back east, then north to the interstate. By this time, both MM were down to about 1/2 tank of gas. Because of the widespread power failures, it was decided to terminate transects and drive west until we could find gas. This was not as easy as we had hoped since power outages continued far into Louisiana. We finally found power and gas in Hammond, Louisiana. Both cars were on or below empty.
The P4 team decided that they would find lodging and rest before returning to Norman the following day. The P3 crew decided that they were rested enough and would drive back to Norman that day. Data collection continued for many hours as we moved through the northwest quadrant of the storm. As we crossed the Mississippi River in Baton Rouge, we ascended a fairly tall bridge (height unknown at this time). Wind speeds increased from about 15 knots to over 35 knots at the high point of the bridge, then decreased again to about 15 knots on the far side. Obviously there was considerable speed shear in this environment.
As we continued northwest, the low overcast gave way to broken and finally to scattered cloud conditions. Data collection was terminated at this time.