DEVELOPMENT OF A LAKE BREEZE FORECAST METHODOLOGY FOR NORTHERN MICHIGAN

Transcription

1 DEVELOPMENT OF A LAKE BREEZE FORECAST METHODOLOGY FOR NORTHERN MICHIGAN Brce B. Smith NOAA/National Weather Service Gaylord, Michigan Abstract It is important for operational forecasters to accrately predict when lake breezes will occr, since lake breezes can have a large impact on temperatre, wind direction, and the development of precipitation. Lake breezes form in response to land/water temperatre differences, which occr freqently on spring and smmer days. The magnitde of the temperatre difference and the strength and direction of the low-level synoptic flow help determine whether or not a lake breeze will form. In this stdy, lake breezes that formed along the Lake Michigan and Lake Hron shorelines in northern lower Michigan from April throgh Agst of 1998 and 1999 were examined. Several variables related to lake breeze formation were evalated, inclding average water temperatre near the shoreline, inland maximm temperatre, the temperatre difference between land and water, forecast 95-mb winds, and whether a lake breeze actally formed. Forecast diagrams were generated showing lake breeze occrrence as a fnction of land / water temperatre dif ference and Eta model forecast 95 mb wind speed. These plots reveal that a relatively light wind is more important than a large land/water temperatre difference when forecasting lake breezes. While lake breezes occrred with a wide range of land / water temperatre differences, lake breeze development tended not to occr when forecast 95 mb wind speeds exceeded approximately 12 to 18 knots, depending on stdy site. These reslts can be sed by operational forecasters to improve forecasts of spring and smmer lake breezes near the Great Lakes. 1. Introdction Lake breezes are common along the Great Lakes dring spring and smmer. These winds form along coastal regions in response to temperatre differences that freqently develop dring the day between the land and water. The lake breeze and its impact on coastal areas has been the sbject of considerable research. Many athors have docmented that the inland penetration of the lake breeze front is accompanied by an abrpt wind shift, decrease in temperatre, increase in relative hmidity and enhanced srface convergence (Estoqe 1962; Frizzola and Fisher 1963; Moroz 1967; Ryznar and Toma 1981). The ability of lake breezes to alter smmertime precipitation patterns has also been noted (Moroz and Hewson 1966). De to the impact lake breezes have on local weather, it is important for opera- tional forecasters to be able to accrately predict lake breeze formation. In this stdy, lake breezes that form over northern lower Michigan along the Lake Hron and Lake Michigan shorelines are examined. The primary goal of the stdy is to develop basic, easily applied gidelines that operational forecasters can se to more accrately predict whether a lake breeze will form, given the prevailing meteorological conditions. Improved lake breeze forecasts shold lead to improved spring and smmer forecasts of temperatre, clods, wind, and precipitation near the Great Lakes. 2. Lake Breeze Characteristics The generally accepted explanation for the development of the lake breeze is depicted in Fig. 1 (adapted from Simpson 1994). When the sn shines dring a typical spring or smmer day, the land qickly becomes warmer, while the water temperatre remains nearly constant. Convection crrents develop over land in response to the rising temperatres, which cases heat near the srface to be redistribted vertically throgh the lowest several thosand feet above the grond. As air over the land is warmed, the air expands and becomes less dense. This prodces a decrease in srface pressre over the land. Meanwhile, the pressre over the water remains nearly constant. The pressre difference (or gradient) between land and water cases air over the water near the shoreline to move inland. This is the lake breeze. Weak sbsidence over the water and a retrn flow aloft directed from land to water completes the lake breeze circlation. Appendix A shows a seqence of srface weather observations associated with the passage of a lake breeze front at Manistee, Michigan (KMBL), located 3 miles from the Lake Michigan shoreline. Note the wind shift and temperatre decrease that occrred at 1835 UTC with the passage of this lake breeze. Lake breeze circlations develop most freqently dring mid morning, several hors after snrise. Lyons (1966), stdying lake breezes on the sotheast shore of Lake Michigan, fond that the lake breeze circlation began at the shoreline arond 8 local time. Moroz and Hewson (1966) fond that the initial onshore flow along the east shore of Lake Michigan started between 9 and 13 local time. Wind speeds associated with lake breezes generally average arond 1 knots. Inland penetration ofthe lake breeze often varies, thogh it can occasionally exceed 3 miles depending on meteorological conditions (Simpson 1994). 47

2 48 National Weather Digest Retrn Flow )././ Low III( < Lake _./ Pressre III( Breeze Wanning Land Lines of Eqal Pressre Cool Water Fig. 1. Conceptal model showing the development of a lake breeze.,-., 7 6 <t>.a 5 4 S' 3 <t> f-; 2 Mean Water/Air Temperatres --- J F M A M J J A SON D Mon th Fig. 2. Mean water temperatre (dashed line) and mean air temperatre over land (solid line) for central Lake Michigan and Mskegon, Michigan, respectively. Mean annal water and air temperatres for central Lake Michigan and Mskegon, Michigan, respectively, are shown in Fig. 2. Note that the annal water temperatre cycle lags the air temperatre cycle. Whereas average air temperatre peaks in mid smmer, average water temperatre does not reach its maximm ntil late smmer. This temperatre lag can be attribted to the difference in heat capacity between land and water, and reslts in abot a five month period (roghly April throgh Agst) when the average air temperatre is higher than the average water temperatre. It is dring this time period (spring and smmer) that lake breezes are most common. In addition to the temperatre contrast between the water and adjacent land, another factor that helps determine whether a lake breeze forms is the strength and direction of the prevailing low level wind. For example, a lake breeze may not develop even with a large land/water temperatre difference if the low level flow is strong and offshore, since this wind wold oppose the onshore wind of a lake breeze. Conversely, a lake breeze may form even with a small land/water temperatre difference provided the opposing offshore low level flow is relatively weak. Frizzola and Fisher (1963) fond that the maximm wind speed that wold permit a sea breeze to form near New York City ranged from 9 to 18 mph. Hall (1954), stdying lake breezes near Chicago, fond that offshore wind speeds of 1 to 15 mph at 2 feet above the srface were the maximm that wold allow lake breeze development. Watts (1955) noted that sea breeze formation on the sothern coast of England depended on both the temperatre contrast and the low level wind direction and speed. He fond that on a calm day, a temperatre difference of 1 C between land and water was large enogh for a sea breeze to form, bt to overcome an offshore wind as strong as 8 m S l (approximately 16 knots) a temperatre difference of 11 C was needed. 3. Methodology This stdy was condcted sing data from April throgh Agst of 1998 and Pertinent data related to lake breeze formation were collected for several sites in northern lower Michigan (Fig. 3). In 1998, the lake breeze stdy sites inclded Alpena, Manistee, and Oscoda, Michigan. In 1999, the stdy sites inclded Alpena, Manistee, and Traverse City, Michigan. Selection ofthese stdy sites was based primarily on their proximity to one of the Great Lakes, and the availability of srface observational data. For each day of the stdy, several variables related to lake breeze development were assessed: Average water temperatre (OC) within approximately 2 miles ofthe shoreline for each site. This information was collected sing the satellite derived water temperatres depicted on the Great Lakes Environmental Research Laboratory's Great Lakes Srface Environmental Analysis. Inland maximm temperatre (OC). De to its proximity to the center of northern lower Michigan, the maximm temperatre at Hoghton Lake (Fig. 3) was considered representative of the inland maximm temperatre for each site. Difference (OC) between the water temperatre and the inland maximm temperatre for each site. Forecast 95 mb wind speed (knots) and direction (to the nearest 1) valid at 15, 18, and 21 UTC for each site. This information was collected from the 12 UTC cycle ofthe Eta model. The existence of a srface front or other srface bondary not associated with a lake breeze in northern Michigan between 12 and UTC. The existence of convection in the vicinity of northern Michigan that may have altered srface winds between 12 and UTC. Observed horly wind direction at each site between 12 and UTC (sed to determine whether an onshore wind associated with a lake breeze occrred). For the prpose of this stdy, a wind was considered onshore if it fell within 7 of a line perpendiclar to the

3 Volme 25 Nmbers 3,4 December Table 1. Percentage of days identified as potential lake breeze days, percentage of potential lake breeze days on which a lake breeze occrred, and the percentage of all days on which a lake breeze occrred. % of Days identified Stdy Site Years as Potential Lake Breeze days Alpena (APN) 1998/99 5% (152/36) Manistee (MBL) 1998/99 3% (92/36) Oscoda (OSC) % (81/153) Traverse City (TVC) % (46/153) % of Potential Lake Breeze days on which Lake Breeze occrred 52% 74% 67% 28% % of All days on which Lake Breeze occrred 26% 22% 35% 8% shoreline for a particlar stdy site. In order to ensre that an onshore wind that developed at a site was associated with an actal lake breeze, only potential lake breeze days were considered. A potential lake breeze day was defined as a day on which all of the following occrred at a stdy site: 1) 95 mb winds were not forecast to be onshore at 15, 18, or 21 UTC 2) A srface front or other srface bondary was not located within northern Michigan between 12 and UTC 3) Convection in the vicinity of northern Michigan was not altering srface winds between 12 and UTC Manistee (MBL) Alpena (APN) Oscoda (OSC) This set of criteria helped ensre that an onshore wind reslting from another meteorological process did not provide the false indication of a lake breeze. Therefore, a lake breeze was considered to have occrred at a stdy site only if the observed srface wind became onshore between 12 UTe and UTe on a day defined as a potential lake breeze day. 4. Reslts and Operational Implications Table 1 shows the percentage of days identified as potential lake breeze days, the percentage of potential lake breeze days on which a lake breeze occrred, and the percentage of all days on which a lake breeze occrred. Note that Oscoda and Alpena had the highest percentage (53% and 5%, respectively) of days identified as potential lake breeze days. This reslt stems from these site's being sitated on Lake Hron's western shore. Here the climatologically favored west wind tends to minimize the nmber of days with an onshore synoptic flow (and therefore increases the nmber of potential lake breeze days). Of those days identified as potential lake breeze days, lake breezes actally occrred most freqently at Manistee and Oscoda (74% and 67%, respectively). This finding can be attribted to the proximity of these sites to their respective lake. Oscoda (2 miles from Lake Hron) and Manistee (3 miles from Lake Michigan) are relatively close to the shore, whereas Alpena (6 miles from Lake Hron) is farther inland. Thogh Traverse City is only 2 miles soth of Lake Michigan, the city borders Grand Traverse Bay. This bay is small (relative to Lake Michigan) and has an irreglarly shaped shoreline, Fig. 3. Sites in Michigan where lake breeze data was collected. Srface observations from Hoghton Lake were sed for the inland maximm temperatre for all stdy sites. " :!;J 3 25 Q 2 c.. g G 15 f-<e a 5...l Alpena (Apr-Ag ) h i! I. f.. :1( :1(..:: I! t.: : h:l( ';$ 'i' ' 'i'!b t ; I '" o mb Wind Speed (knots) Fig. 4. Scatterplot showing the combination of land/water temperatre difference and Eta model forecast 95 mb wind speeds (valid at 18 UTC) reslting in days with lake breezes (dots) and days withot lake breezes (stars) at Alpena, Michigan. Dashed line represents sbjective delineation between days with lake breezes and days withot. Only potential lake breeze days are plotted. Total sample size is 152. sggesting that lake breezes that form at Traverse City are likely more complex than those that form at the other stdy sites. Overall, lake breezes occrred most freqently at Oscoda (35% of all days) and Alpena (26% of

4 5 National Weather Digest " 3 25 is E-... " 1 '" : 5...J '"" Manistee (Apr-Ag ). I. I I. I I.I. li: I I "'. I. f: :t>.-!..,.. '", o mb Wind Speed (knots) Fig. 5. Same as in Fig. 4., except for Manistee, Michigan. Total sample size is " 25 Oscoda (Apr-Ag 1998) is ' E t:) '" " 1 : 5...J "'". ' '" II.,..,., JI( f ". -.!.!. '" o mb Wind Speed (knots) Fig. 6. Same as in Fig. 4., except for Oscoda, Michigan. Total sample size is " 25 is E-... " 1 '" 't:l 5...J Traverse City (Apr-Ag 1999) I. ' :. t li: :,. --/:i:' o mb Wind Speed (knots) Fig. 7. Same as in Fig. 4., except for Traverse City, Michigan. Total sample size is 46. all days), and least freqently at Traverse City (8% of all days). Figres 4-7 show lake breeze occrrence and nonoccrrence for each stdy site as a fnction of land/water temperatre difference and Eta forecast 95 mb wind speed valid at 18 UTC. These figres inclde potential lake breeze days only. In each figre, a dot indicates that a lake breeze occrred on a potential lake breeze day, while a star indicates that a lake breeze did not occr on a potential lake breeze day. The crved, dashed line in each figre delineates the combinations of land/water temperatre differences and forecast 95 mb wind speeds that are favorable for lake breezes (left of the crved line) from those that are not favorable for lake breezes (right of the crved line). These ''best fit" crved lines were drawn sbjectively and assme that a positive temperatre difference (land warmer than water) is reqired to generate a lake breeze. The fact that some data points fall on the wrofig side of the ''best fit" lines can be attribted to a nmber of factors, inclding inaccrate inland maximm temperatre, inaccrate assessment of water temperatre, inaccrate 95 mb forecast wind speed, and other small scale weather interactions (like mesoscale bondaries and terrain indced winds). As previosly noted, both the contrast in temperatre between land and water, and the strength and direction of the prevailing low level synoptic flow help determine whether a lake breeze forms. Conceptally, a light wind speed and/or a large temperatre contrast shold favor the development of a lake breeze. Figres 4-7 sggest, however, that a relatively light (forecast) 95 mb wind speed is more important than a large land / water temperatre contrast when forecasting lake breeze development. Note that lake breezes occrred at each stdy site with a broad range of land/water temperatre differences. With the exception of Traverse City, lake breezes even occrred with a temperatre contrast as low as 2-3 C. As the forecast 95 mb wind speed increased, however, lake breeze freqency at each site showed a clear decrease. Figres 4-7 sggest that the maximm forecast 95 mb wind speed associated with a lake breeze ranged from approximately 12 knots at Traverse City to abot 18 knots at Manistee. These vales are consistent with the findings of Frizzola and Fisher (1963), Hall (1954), and Watts (1955). With forecast responsibility along portions of Lake Sperior, Lake Michigan, and Lake Hron, meteorologists at the National Weather Service (NWS) in Gaylord, Michigan, rotinely consider lake breezes when issing several types of forecasts dring spring and smmer. Forecasters at NWS Gaylord reference Figs. 4-7 when deciding on the likelihood oflake breezes, and have fond the diagrams to be a valable sorce offorecast gidance. When lake breezes are anticipated, forecasters coordinate and reference the reslting onshore wind in all appropriate forecasts (Zone Forecasts, Nearshore Marine Forecasts, and Aviation Forecasts). Appendix B shows examples of these NWS forecasts on a day when lake breezes are expected. 5. Smmary Lake breezes are common along the shores of the Great Lakes dring spring and smmer. These circlation patterns develop in response to land/water temperatre gradients that form when the land is warmer than the adjacent water. Whether or not a lake breeze forms is a fnction of the magnitde of the land/water temperatre difference, and the strength and direction of the low level synoptic flow.

5 Volme 25 Nmbers 3,4 December In this stdy, lake breezes that formed along the Lake Michigan and Lake Hron shorelines dring April throgh Agst of 1998 and 1999 were examined. Stdy sites inclded Alpena, Manistee, Oscoda, and Traverse City, Michigan. The goal of the stdy was to develop gidelines that operational forecasters cold easily se to more accrately predict the occrrence of lake breezes. For each stdy site, several variables related to lake breeze development were assessed, inclding average water temperatre near the shoreline, inland maximm temperatre, the temperatre difference between land and water, forecast 95 mb winds, whether convection and/or bondaries were affecting northern Michigan, and observed wind directions. Forecast diagrams were generated for each stdy site showing how lake breeze occrrence depended on land/water temperatre difference and forecast 95 mb wind speed. These plots indicated that a relatively light (forecast) 95 mb wind speed was more critical than a large land/water temperatre difference when forecasting lake breeze development. While lake breezes were fond to occr within a broad range of land/water temperatre differences, an pper limit of forecast 95 mb wind speeds was fond to be associated with lake breeze development. This pper limit ranged from approximately 12 to 18 knots, depending on stdy site. Reslts from this stdy can be sed by operational forecasters to improve forecasts of spring and smmer lake breezes near the Great Lakes. Improved lake breeze forecasts will sbseqently lead to improved forecasts oftemperatre, clod, wind, and precipitation in the vicinity of the Great Lakes. Acknowledgments The athor wold like to thank NWS forecasters Scott Rozanski and Tom Spriggs for assisting with the data collection dring this stdy. Also, reviews by NWS regional science officer Preston Leftwich, NWS forecaster Steve Considine, and two additional reviewers led to significant improvements in the qality of the manscript. Athor Brce Smith is the Science and Operations Officer (SOO) at the NWS Forecast Office in Gaylord, Michigan. He is responsible for local forecast techniqe training and development, and for overseeing onstation research activities. He received his B.S. and M.S. in Atmospheric Science from The University of Michigan in 1989 and 199, respectively. Prior to becoming the SOO at NWS Gaylord in 1995, he was a Meteorologist Intern at NWS Lansing, Michigan, from 199 to 1992; a Jorneyman Forecaster at NWS DetroitIPontiac, Michigan, from 1992 to 1994; and a Lead Forecaster at NWS DetroitIPontiac, Michigan, from 1994 to His research interests inclde lake effect snow, model verification, and the impact of terrain and land/water interactions on local weather conditions. References Estoqe, M. A., 1962: The sea breeze as a fnction of the prevailing synoptic sitation. J. Atmos. Sci., 19, Frizzola, J. A., and E. L. Fisher, 1963: A series of sea breeze observations in the New York City area. J. Appl. Meteor., 24, Appendix A Seqence of srface weather observations from Manistee, Michigan, associated with the passage of a lake breeze: METAR KMBL 91335Z AUTO 147KT 1SM FEW17 FEW41 SCT55 14/1 A2966 RMK A1 METAR KMBL 91435Z AUTO 148KT 1SM FEW95 16/9 A2966 RMK A1 METAR KMBL 91535Z AUTO 124KT 1SM FEW11 15/8 A2968 RMX A1 METAR KMBL 91635Z AUTO 125KT 1SM CLR 16/8 A2968 RMK A1 METAR KMBL 91735Z AUTO 129KT 1SM CLR 16/8 A2968 RMK A1 METAR KMBL 91835ZAUTO 2912G19KT 9SM SCT38 13/7 A2964 RMK A1 METAR KMBL 91935Z AUTO 38KT 8SM SCT1 SCT33 OVC41 13/8 A2971 RMK A1 METAR KMBL 9235Z AUTO 319KT 7SM SCT1 BKN A2971 RMK A1 METAR KMBL 92135Z AUTO 39KT 1SM FEW9 13/9 A2968 RMK A1 METAR KMBL 92235Z AUTO 179KT losm FEW75 FEW A2966 RMK A1 Appendix B Excerpts of pblic, marine, and aviation forecasts issed by NWS Gaylord, Michigan, on a day when lake breezes were expected: Zone Forecast Prodct (ZFP):.TODAY... SUNNY AND WARM. HIGHS NEAR COOLER NEAR LAKE HURON. WEST WINDS AROUND 1 MPH BECOMING ONSHORE. Nearshore Marine Forecast (NSH):.TODAY... WEST WINDS AROUND 1 KNOTS BECOMING ONSHORE. SUNNY. WAVES ONE FOOT OR LESS. Aviation Forecast (TAF): KAPN Z KT P6SM SKC BECMG KT FM17 111KT P6SM FEW6 FMOOOO 27KT P6SM SKC BECMG 23 VRB3KT =

6 52 National Weather Digest Hall, C. D., 1954: Forecasting the lake breeze and its effects on visibility at Chicago Midway Airport. Bll. Amer. Meteor. Soc., 35, Lyons, W. A., 1966: Some effects of Lake Michigan pon sqall lines and smmertime convection. Great Lakes Res. Div. Pb., 15, Moroz, W. J., 1967: A lake breeze on the eastern shore of Lake Michigan: Observations and model. J. Atmos. Sci., 24, Ryznar, E., and J. S. Toma, 1981: Characteristics of tre lake breezes along the eastern shore of Lake Michigan. Atmos. Environ., 15, Simpson, J. E., 1994: Sea breeze and local wind. Cambridge University Press, 234 pp. Watts, A. J., 1955: Sea breeze at Thomey Island. The Meteorological Magazine, 84, , and E. W. Hewson, 1966: The mesoscale interaction of a lake breeze and low-level otflow from a thnderstorm. J. Appl. Meteor., 5, ABOUT OUR CORPORATE MEMBERS Smmaries are provided from corporate members to inform all readers of the wide variety of activities operational meteorologists and those in related fields are involved in. STORM STOCK Fonded in 1993, Prairie Pictres' StormStock library is recognized as the highest qality collection of storm footage in the world. StormStock images are licensed to film and television indstry prodcers for se in broadcast, theatrical, advertising and corporate video prodctions. "What makes StormStock so sccessfl is or commitment to qality," says fonder and cinematographer Martin Lisis. "We've established a significant collection of storm footage on premim formats like 35mm, for instance. We have images that don't exist anywhere else," he said. Dedication to image qality, combined with otstanding cstomer service, frthers StormStock's top ranking among TV and film prodcers arond the globe. Campbell-Ewald Advertising prodcer Craig MacGowan was pleased with StormStock saying, "When the creative team at Campbell-Ewald came to me with their Chevrolet ' Storm Chaser' commercial for Chevy S-1 pickp trcks, I knew I had to go the extra mile for them on the prodction. That meant I had to find the highest qality, rock-steady stock footage I cold find. Time and again, reel after reel of stock footage, and the Prairie Pictres [StormStock] footage was the clear leader. "When prodcers contact s, they have a concept in mind for what kind of stock they need," Lisis said. "We can qickly detelmine what will work best for them becase we know or library in and ot. We get the right images to or clients in a short period of time. I think or sccess, overall, is de largely to the fact that shooters operate StormStock. Clients appreciate the opportnity to deal directly with a cinematographer," he said. Lisis admits that tracking and photographing storms is the most exciting part of the job, "We've seen or share of incredible storms over the years," he said. "We've captred tornadoes, driven throgh baseball hail and 1 mph winds, watched 1- mile high spercell thnderstorms bild and twist like barbershop poles, and dodged white-hot c1od-to-grond lightning. Confronting severe weather is scary to most people, bt it's jst another day at the office for s." StormStock can be accessed on-line at The Web site offers sample video clips from the library, and ordering information. Storm Stock clients inclde ABC News, Bozell Worldwide, ChiatlDay, CBS News, CNN. Dateline NBC, Discovery Channel, HBO, Indstrial Light & Magic, Leo Brnett USA, National Geographic Television, The X-Files, and others. For more information, contact them at: Prairie Pictres' StormStock, P.O. Box 1222, Arlington, Texas 7612 Tel: ; info@stormstock.com. ADVANCED DESIGNS CORPORATION Advanced Designs Corporation (ADC), the only athorized domestic Collins radar system provider, now over twenty years in the weather radar and weather data display indstry, continally adds featres and options to enhance the capabilities of or prodcts. ADC is committed to having the finest radar and display prodcts available for Collins radars. We were first to develop a broadcast display nit for Collins grond-based radars. ADC delivered the first flly coherent Doppler radar for broadcast back in ADC's in-hose hardware/software development team keeps s on the ctting edge of new radar and display technologies. Whether yo are concerned with lightning, mean radial velocity, storm tracking, reflectivity and Doppler trblence, and/or distribting radar data in real-time to any nmber of sbscribers, ADC has a prodct that meets yor needs. Or single compter PC-based systems are proven reliable and ser-friendly. Collins and ADC work together to create systems that increase the capabilities of the end-ser, many of which are proprietary in natre and will be available only to or cstomer. ADC can provide Doppler radar soltions for the offshore oil indstry, the military, tility companies, governments, aerostats, the broadcast indstry, or any other application where monitoring severe weather is necessary. Mobile configrations can also be provided. For more information, contact ADC at 1169 West 2nd Street, Bloomington, Indiana 4743; Tel: ; Fax: ; Internet: adc@doprad.com