An Approach to Waterspout Forecasting for

South Florida and the Keys

Daniel P. Brown and Joel Rothfuss

National Weather Service Forecast Office, Miami, FL


Waterspouts are extremely common during the rainy season over the coastal waters of the Florida Keys and the southeast Florida Coast. Rarely are waterspouts life threatening, but they occasionally move onshore and produce minor structural damage. Waterspouts can pose a threat to recreational boaters and coastal residents.


Dr. Joe Golden (1970, 1973, 1974a, 1974b, 1977) completed several studies and papers on the formation, life cycle and frequency of waterspouts in the Florida Keys. Additional studies (Clemons, 1969 and Garrish, 1967) have been published which discussed the climatology of waterspouts along the southeast Florida coast and Keys.
After reviewing several of these papers it may be possible to develop a forecasting scheme to determine certain days which are more favorable for waterspout formation. Special Marine Warnings are usually issued after a public or pilot report is received and therefore is a hind cast and not a forecast. The ideas presented in this paper could radically change this perception and possibly increase the awareness of days in which waterspout formation is favorable. Choy and Spratt (1994) stated that it would be useful to heighten awareness of coastal residents and the marine community before a potential waterspout event. If this can be implemented, highlights in the South Florida Hazardous Weather Outlook, Nowcasts, and Marine Weather Statements could possibly be included to alert boaters and residents of the daily waterspout threat.


Golden (1977) found that waterspouts occur more frequently in the Florida Keys than anywhere else in the world. He discovered that between 50 and 500 waterspouts occur each year in the Keys. In 1969, at least 400 waterspouts were documented by Golden (1970) during the Lower Keys Waterspout Project. The second most active area is along the southeast Florida coast from Stuart to Homestead.
Most waterspouts are relatively weak, however, Golden (1977) has measured maximum rotational winds as high as 190 knots in Florida Keys waterspouts. Golden (1973) found that only 30 percent of waterspouts have associated thunder and lightning. Golden (1974a), along with many others, have noted that funnel clouds over water should be treated as waterspouts, as in weak waterspouts, the visible funnel does not often extend to the sea surface even though sea spray is picked up.


A study by Clemons (1969) and additional studies by Golden (1970) revealed similar results in Key's waterspout climatology. Both concluded that waterspout activity rises sharply in late April to mid-May, reaches a relative maximum in June, then falls in July before reaching its primary peak in August. The activity remains high in September before sharply decreasing in October.
Golden (1970) found that days with weak low-level vertical shear, without any well organized synoptic scale disturbances, were ideal for waterspout development. Since the prevailing wind direction is southeast, Clemons found that 66% of the events occurred with wind directions, of east through south. The highest prevailing wind speed recorded at Key West during waterspout formation was 16 knots with speeds between seven and 10 knots recorded during 57% of the events.
Clemons found that the maximum waterspout activity in the Lower Keys is during the noon hour with a secondary maximum during the late afternoon. The peak activity times were slightly later than Gerrish (1967) found for the Southeast Florida area. Golden (1974b) also noted that most waterspouts commonly form in lines of building cumulus clouds. Only about 5 percent of waterspouts occur in isolated cumulus clouds. Most waterspouts occur during times of benign weather and tend to occur on successive days. Clemons found that two to three days in a row with waterspout formation is common with the maximum of six consecutive days in June of 1968.


The purpose of developing a forecasting scheme is to try to determine days which are more favorable for waterspout formation. Using the studies on waterspout climatology by Clemons, Gerrish, and Golden, a simple forecasting worksheet was produced. The forecasting sheet weighted certain wind directions and speeds higher than others. Mean wind speeds and directions below 600mb (approximately 14,000 feet) were estimated by forecasters using the 1200 UTC Miami and Key West Soundings. These mean wind speed and direction calculations were then figured into a simple calculation sheet to determine the waterspout potential. The forecasting scheme also included a persistence factor which used the notion that waterspout formation usually occurs on successive benign weather days. The formula used to compute favorable waterspout days follows:
1.) Mean wind below 600mb less than 8 kt, with no wind above 12 kt...... add 1.0
Mean wind below 600mb between 8 and 11kt with no wind about 14kt..add 0.5 -------
2.) Mean wind direction below 600mb
Between 080-120 degrees.......................add 0.5
Between 120-160 degrees.......................add 1.0
Between 160-180 degrees.......................add 0.5
Between 210-360 degrees and 0-70 degrees..................subtract 0.5 -------
3.) Previous day had waterspout activity and no synoptic changes expected...add 1.0 -------

If total equals 2.0 or greater, waterspouts are likely today. --------

One limitation of the forecasting scheme is that it did not take into account any stability or moisture parameters. Since only minor airmass changes occur in south Florida during the summer, especially in the low levels, these parameters were not used. It was noted that most waterspout days had precipitable water amounts of 1.8 inches or greater. However, waterspouts were observed on at least three days in the Keys with precipitable water amounts between 1.4 and 1.5 inches. More studies will need to be completed so that moisture and stability parameters can be used for the waterspout forecasting scheme. Until more studies are completed, it should be the forecasters' responsibility to determine the likelihood of showers over the coastal waters and factor this into their final assessment of the waterspout potential.
Another aspect of the scheme which may be more important than the moisture or stability parameters are whether any cloud lines or boundaries are present over the coastal waters. These boundaries may include land breeze boundaries which sometimes form during the overnight hours along the Southeast Florida coast. In the Keys cumulus cloud lines typically form in the late morning and early afternoon on the lee side of the larger islands. Additionally, outflow boundaries from previous afternoon convection can sometimes remain over the South Florida coastal waters. These boundaries seem to be a focus for coastal showers which typically produce waterspouts. Currently, the boundary factor is not directly included in the calculation. But, if the calculation sheet determines that waterspouts are likely and a boundary is present on satellite pictures, then it is recommended that a Marine Weather Statement be written describing the position of the boundary and that the likelihood of waterspouts along the boundary will be high.



Specific results were computed using the months of June through September. Days on which 1200 UTC sounding at Miami or Key West was not available were deleted from the computation.Of the 16 waterspout days along the southeast Florida Coast in which a Miami 1200 UTC sounding was available, the forecasting scheme correctly predicted 6 of these days or 38 percent. In the Keys of the 30 waterspouts days with a 1200 UTC Key West sounding 14 were correctly predicted or 47 percent. The percentage of days in which waterspouts were predicted and actually occurred was around 45 percent for both the southeast Florida coast and Keys. The forecasting scheme missed slightly more than half of the days in which waterspouts were reported during the 1997 waterspout season. Overall, the use of the forecasting scheme enabled forecasters to predict about half of the waterspout days during the 1997 season. Even though higher detection rates and lower false alarm rates were hoped for, the results do indicate an improvement in waterspout forecasting by the NWSFO in Miami.
Climatology of the 1997 waterspout season was also computed using the 1200 UTC soundings from Key West and Miami. Data from the surface to 600mb (14,000 feet) from Key West indicated that 75 percent of the waterspouts occurred with prevailing mean wind directions of east to south (Table 1). Eleven percent occurred respectively from both a north to northeast and southwest to west wind directions. These findings are similar to the previous studies of Golden and Clemons. Using the Miami 1200 UTC sounding, similar results occurred along the Southeast Florida coast. Seventy percent of the waterspouts occurred with the mean wind direction between east and south (Table 2). However, a surprising result was noted with 19 percent of the spouts occurring from southwest to west wind direction. These days were most likely days in which waterspouts were associated with afternoon showers and thunderstorms moving off the southeast coast. The remaining 11 percent occurred with a north to northeast prevailing wind direction.

Table 1. Prevailing Mean Wind Direction Below 600mb (approx.14,000 ft.)
For the Florida Keys Waterspout Days
From the 1200 UTC Sounding Data from Key West
Mean Wind Direction Number of Waterspouts Days Percentage of Total
N to NE (00 to 60 degrees) 4 11%
E to S (70 to 200 degrees) 27 79%
SW to W (210 to 290 deg.) 4 11%
NW to N (300 to 360 deg.) 1 3%

Table 2. Prevailing Mean Wind Direction Below 600mb (approx.14,000 ft.)
For the Southeast Florida Coast Waterspout Days
From the 1200 UTC Sounding Data from Miami
Mean Wind Direction Number of Waterspouts Days Percentage of Total
N to NE (00 to 60 degrees) 3 11%
E to S (70 to 200 degrees) 19 70%
SW to W (210 to 290 deg.) 5 19%
NW to N (300 to 360 deg.) 0 0%

This study also found that light winds favor waterspout formation. Seventy-two percent of the waterspouts in the Keys occurred with mean wind speed 8 knots or less (from the 1200 UTC Key West sounding) (Table 3). Twenty-five percent occurred with mean wind speeds between 9 and 12 knots, with only one spout occurring with mean wind above 12 knots. Similar results were noted along the Southeast Florida coast as 67 percent of the waterspouts occurred with wind speed 8 knots or less. Twenty-six percent occurred with mean wind speeds between 9 and 12 knots and 7 percent occurred with speeds greater than 12 knots (Table 4).
During the 1998 waterspout season seven times waterspouts occurred on successive days in the Keys. Along the Southeast Florida Coast successive waterspout days occurred four time. The number of days in a row with waterspout was five consecutive day in the Keys from July 9 through the 13. As previous studies have pointed out, successive waterspout days are quite common.
Table 3. Prevailing Mean Wind Speed Below 600mb (approx.14,000 ft.)
For the Florida Keys Waterspout Days
From the 1200 UTC Sounding Data from Key West
Mean Wind Speed Number of Waterspouts Days Percentage of Total
Less than or equal to 8 knots 26 72%
Between 9 and 12 knots 9 25%
Greater than 12 knots 1 3%

Table 4. Prevailing Mean Wind Speed Below 600mb (approx.14,000 ft.)
For the Southeast Florida Coast Waterspout Days
From the 1200 UTC Sounding Data from Miami
Mean Wind Speed Number of Waterspouts Days Percentage of Total
Less than or equal to 8 knots 18 67%
Between 9 and 12 knots 7 26%
Greater than 12 knots 2 7%
The forecasting scheme was able to increase the forecaster awareness of waterspout potential and improved the timeliness of statements issued by the National Weather Service Forecast Office in Miami. However, the forecasting scheme will continue to be updated and slight modifications are expected to be incorporated in 1998.


Based on findings from the 1997 waterspout study, a new calculation sheet will be used during the upcoming 1998 waterspout season. The new calculation sheet will use slightly different calculations for wind directions, but the wind speed and persistence calculations will remain the same. The new calculation sheet has also increased the final numbers needed to issue statements, however, it remains the forecasters discretion when to issuing statements concerning the daily waterspout potential. The updated calculation sheet is included in Table 5.
It is hoped that during the 1998 waterspout season, a new data source can be used to compute the daily mean wind speed and direction. Between the 1997 and 1998 waterspout seasons aircraft sounding data called ACARS (1997) was made available to the Miami NWSFO through the World Wide Web. ACARS data is available from selected aircraft departing and arriving at Miami and Palm Beach International Airports. Since the data from the Miami 1200 UTC sounding is not available until approximately 9:00 A.M., the ACARS sounding data is an excellent supplement to the 1200 UTC soundings. The ACARS data includes wind speed and direction in both a text and graphical format. A simple computer program written by Carroll (1998) of the NWSFO in Miami will be used to compute the mean wind speed and direction from the ACARS data to aid in the calculation of the waterspout forecasting scheme. It is believed that with the supplemental use of ACARS data, improvements in forecasting waterspout days can be achieved.
The results of the 1997 waterspout study indicated that heightened forecaster awareness and the use of the forecasting scheme will make it possible to highlight favorable waterspout days in statements issued by the Miami NWSFO. In most cases, Special Marine Warnings continued to be issued after the initial report of waterspouts. However, statements such as the South Florida Hazardous Weather Outlook, Nowcasts, and Marine Weather Statements were used to mention the possibility of waterspout formation. Marine Weather Statements and Nowcasts often gave several minutes to a few hours lead time. The South Florida Hazardous Weather Outlook often gave a "heads-up" to the marine community several hours in advance of any waterspout formation. The forecasting scheme was not intended to correctly forecast every single waterspout day but was intended to highlight days in which confidence of waterspout formation is high.
Hopefully, further studies using additional meteorological parameters can be completed to narrow down the most favorable conditions for waterspout formation. As the 1997 waterspout study is extended to the 1998 waterspout season, it is believed that the goal of higher detection and lower false alarms rates can be achieved. As these goals are achieved, heightened awareness of boaters and coastal residents can help the National Weather Service to achieve its main goal of saving both lives and property in South Florida.

Table 5: Calculating Waterspout Potential

Note: The waterspout calculations do not include stability or moisture parameters. Before using make sure stability and moisture parameters are high enough to support at least widely scattered showers over the coastal waters.
ACARS / Sounding mean wind speed and direction______________

1.) If mean wind below 15,000 feet is (equal to or less than) 8 knots
and max layer wind speed 12 kt or below then.........................add 1.0
(If max layer wind is between 12 kt and 15 kt then only add 0.5)
If mean wind below 15,000 feet is between 9 and 11 kt and
max layer wind speed 15 kt or less then............................add 0.5
(If max layer wind above 15 kt then add 0)

If mean wind below 15,000 feet is greater 12 kt or greater then..add 0
Wind Speed Factor_________
2.) If mean wind direction below 15,000 feet is:

Between 010-069 degrees.............................add 0.5
Between 070-209 degrees.............................add 1.0
Between 210-289 degrees.............................add 0.5
Between 290-360 degrees.......................subtract 0.5

Wind Direction Factor_________
3.) If previous day had waterspout activity and no major synoptic
change expected then.............................................................add 1.0
(Treat waterspouts in Keys and SE Coast Seperately)
Persistence Factor_________

If total equals 2.0 then waterspouts are possible and may be highlight in HWO.
If total equals 2.5 or 3.0 then waterspouts potential is high and should be highlighted in HWO or other statements

Note: If confidence is high and convergent cloud line is present on satellite or radar, or is expected to develop over the coastal waters, a MWS describing the position of the line and the high likely of waterspouts is recommended.


Carroll, S. E., 1998: ACARS Generated Helicity Computer Program, Miami, FL, 1998.

Choy, B.K., and S.M. Spratt, 1994: NOAA Tech. Memo. NWS SR-156, 24pp.

Clemons, G. H., 1969 : Detailed Radar Observations and Recent Climatology of Waterspouts in the Key West Area, Preprints Sixth Conf. Severe Local Storms, Chicago, Ill., Amer. Meteor. Soc., 172-175.

Gerrish, H. P., 1967: Tornadoes and Waterspouts in the South Florida Area, Proc. of the 1967 Army Conference on Tropical Meteor., Coral Gables, Fla., June 8-9, 1967.

Golden, J. H., 1970: The Lower Florida Keys Waterspout Project, May-September 1969. Bull. Amer. Meteor. Soc., 51, 235-236.

----, 1973: Some statistical aspects of waterspout formation. Weatherwise, 26, 108-117.

----, 1974a: Life cycle of Florida Keys' waterspouts I. J. Appl. Meteor., 13, 676-692.

----, 1974b: Scale-Interaction Implications for the Waterspout Life Cycle. II. J. Appl. Meteor., 13, 693-709

----, 1977: An Assessment of Waterspout Frequencies along the U.S. East and Gulf Coasts, J. Appl. Meteor., 16, 231-236.

NWS Forecaster's Guide to ACARS Aircraft Data, Washington, 1997. is the U.S. government's official web portal to all federal, state and local government web resources and services.