Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Item L1
BOARD OF COUNTY COMMISSIONERS AGENDA ITEM SUMMARY Meeting Date: November 17, 2010 Division: _Grov,th Management Bulls Item: Yes _ No x Department: Staff Contact Person/Phone #: Christine Hurlev, 289-2517 AGENDA ITEM WORDING: Discussion of Monroe County Hurricane Evacuation Clearance Time Final Report, prepared by Reid ENving, Ph.D. dated November 8, 2010. ITEM BACKGROUND: Objective 101.2 of the Monroe County Comprehensive Plan requires Monroe County to reduce hurricane evacuation clearance times to 24 hours by the year 2010. In November, 2009, County staff and other State of Florida agencies attended a Hurricane Modeling Work group meeting to develop various assumptions (see appendix of report) to be used in the hurricane evacuation modeling. Since development of the initial (MILLER) model used previously for hurricane evacuation modeling, the County also amended the comprehensive plan, adding Policy 216.1.8, ii--hick requires phased evacuation ii--ith visitors leaving 48 hours, mobile homes leaving 36 hours, and permanent residents leaving 30 hours in advance of tropical storm ii--inds in a category 3-5 storm. Tii,-o "assumptions" used previously in the modeling effort have been substantially changed since the initial model ii-as developed including participation rates and floe rates: 1) 70% participation (meaning 70% of people ii--ould evacuate) and 2)maximum FDOT capacity of roadiiays. The DCA retained Dr. Baker to sm-N-ey approximately 400 residents in unincorporated Monroe County and his data indicated closer to 90% of those surveyed ii--ould evacuate. FDOT provided updated floii rates on June 18, 2010. Dr. Eii,-ing has also added a scenario in his report that includes the FDOT 5-Year Work Program roadie --aye projects for the Florida Keys Emergence Evacuation Improvements (outbound functional evacuation lanes) that ii,-ere developed in the 2001 Florida Keys Hurricane Evacuation Study Table 18. Please note that in 2008, the BOCC requested a 4' paved shoulder in lieu of a 10' paved shoulder for projects bete,-een MM93- 106 in the 5-Year Work Program; hoii-ever, the 5-Year Work Program still includes the segments for milling and resurfacing and the 10' shoulder has been included in the model results beloii because FDOT is aii,-aiting the DCA final decision on modeling assumptions. Dr. Eii-ing also added a scenario that demonstrates evacuation time ii--ithout the outbound shoulder betii--een MM90-106 being used as evacuation lanes. The November 8, 2010 final report includes clearance times and data from the neii phased model that includes results for a range of participation (70%-75% and 90%-95%). It also includes the initial floe rates recommended for each road segment and neii floe rates recommended for each road segment submitted by FDOT. Finally, it includes clearance times and results if FDOT makes all improvements in the 2015 Work Program (and if all improvements are made except shoulder ii--idening betii--een MM9-106), evacuation to Florida City, as opposed to the FI J Shelter, occupancy rates based on today's unit occupancy estimates and occupancy based on a more conservative assumption, and the start time for evacuation starts at 30 hours. The results are as folloll--s: Clearance Times (relative to the permanent unit evacuation order) Loii,- Occupancies (2001) High Occupancies (2008) Occupancy by Zone 1=67%; 2=54%; Occupancy by Zone 1=84%; 2=67%; 3=47%; 4=35%; 5=46%; 6=52%; 3=59%; 4=44%; 5=58%; 6=65%; 7=27% 7=34% Loii,- High Participation Loii,- Participation High Participation Participation Approx 90-95% Approx 70% Approx 90-95% Approx 70% 2001 Lanes/2001 Miller 16 hours 16 18 hours 50 18 hours 32 22 hours 6 Flory Rates minutes minutes minutes minutes 2001 Lanes/2010 FDOT 18 hours 58 22 hours 28 22 hours 8 27 hours 2 Flory Rates minutes minutes minutes minutes 2015 Lanes/2010 FDOT 16 hours 16 16 hours 16 16 hours 16 18 hours 40 Flory Rates minutes minutes minutes minutes 2015 Lanes/2010 FDOT 16 hours 16 17 hours 16 17 hours 4 20 hours 16 Floe-- Rates (ii--ithout minutes minutes minutes minutes outbound shoulder from min 90 to min 106) The next step is for the Department of Community Affairs to submit the Annual Report to the Cabinet, with Hurricane Evacuation Report, by the end of November. As you know, ROGO allocations are tied to hurricane evacuation. It is possible annual ROGO allocations may change, pending State review and acceptance of the various assumptions in the report. PREVIOUS RELEVANT BOCC ACTION: At the October 20, 2010 BOCC meeting the Board asked that the item be brought back to them at the November 17, 2010 meeting. CONTRACT/AGREEMENT CHANGES:n/a STAFF RECOMMENDATIONS: TOTAL COST: n/a INDIRECT COST: BUDGETED: Yes No _ DIFFERENTIAL OF LOCAL PREFERENCE: COST TO COUNTY: SOURCE OF FUNDS: REVENUE PRODUCING: Yes _ No AMOUNT PER MONTH Year APPROVED BY: County Atty OMB/Purchasing Risk Management DOCUMENTATION: Included x Not Required _ DISPOSITION: AGENDA ITEM 9 MEMORANDUM MONROE COUNTY GROWTH MANAGEMENT DIVISION We strive to be caring, professional and fair To: Monroe County Board of County Commissioners From: Christine Hurley, AICP Division Director Date: November 2, 2010 RE: Hurricane Evacuation Modeling As you know, Objective 101.2 of the Monroe County Comprehensive Plan requires Monroe County to reduce hurricane evacuation clearance times to 24 hours by the year 2010. In addition, Monroe County amended their Comprehensive Plan in 2005 to require phased evacuation as follows: Policy 216.1.8 In the event of a pending major hurricane (category 3-5) Monroe County shall implement the following staged/phased evacuation procedures to achieve and maintain an overall 24-hour hurricane evacuation clearance time for the resident population. 1. Approximately 48 hours in advance of tropical storm winds, a mandatory evacuation of non-residents, visitors, recreational vehicles (RV's), travel trailers, live-aboards (transient and non -transient), and military personnel from the Keys shall be initiated. State parks and campgrounds should be closed at this time or sooner and entry into the Florida Keys by non-residents should be strictly limited. 2. Approximately 36 hours in advance of tropical storm winds, a mandatory evacuation of mobile home residents, special needs residents, and hospital and nursing home patients from the Keys shall be initiated. 3. Approximately 30 hours in advance of tropical storm winds, a mandatory phased evacuation of permanent residents by evacuation zone (described below) shall be initiated. Existing evacuation zones are as follows: a) Zone 1— Key West, Stock Island and Key Haven to Boca Chica Bridge (MM 1-6) b) Zone 2 — Boca Chica Bridge to West end of 7-mile Bridge (MM 6-40) c) Zone 3 — West end of 7-Mile Bridge to West end of Long Boat Key Bridge (MM 40- 63) d) Zone 4 — West end of Long Boat Key Bridge to CR 905 and CR 905A intersection (MM 63-106.5) e) Zone 5 — 905A to, and including Ocean Reef (MM 106.5-126.5) The actual sequence of the evacuation by zones will vary depending on the individual storm. The concepts embodied in this staged evacuation procedures should be embodied in the appropriate County operational Emergency Management Plans. The evacuation plan shall be monitored and updated on an annual basis to reflect increases, decreases and or shifts in population; particularly the resident and non-resident populations. [9J-S. 012(3)(c)4] The attached "Work Program Requirements Pursuant to Rule 28-20.110 FAC Hurricane Evacuation Carrying Capacity Implementation Year 8, Task Q (July 12, 2005) table shows the excerpt from the State of Florida WORK PLAN related to Hurricane Evacuation modeling to determine clearance times as of 2010. As you can see, the work plan calls for DCA to enter into a Memorandum of Understanding (MOU) with County and City governments for variables and assumptions to be used for the Evacuation Modeling. In November, 2009, DCA convened a work group of professionals to go over certain assumptions used in the 2001 Hurricane Evacuation Miller Model, compared to those assumptions professionals in this field had studied since 2001(after various hurricane evacuations in other areas and surveys that had been conducted). DCA developed a matrix of the various assumptions (see pages 26-28 of 2nd attachment). After the work group convened in November, 2009, other professionals (including DCA consultant [Baker] and FDOT consultants [Wolshon and Joaquin Vargas]) also developed alternative assumptions (primarily flow rates for vehicles based on direct observation and events in the Florida Keys and other geographic areas) throughout 2010 (see memorandum dated June 18, 2010 from FDOT). As we were nearing the time when we had to submit our annual "report" to DCA on progress that had been made on work plan items and since assumptions had not been agreed upon via MOU, the County Consultant, Dr. Reid Ewing, was asked to go ahead and publish his final report, absent the agreement on assumptions. That is why the report has different assumptions for participation rates and for different flow rates (roadway capacities) results with the different variables found on page 24. Dr. Ewing was originally retained to edit the Hurricane Evacuation Model, primarily to add phasing to the model consistent with the County's Comprehensive Plan Amendment done in the mid 2000's that calls for phased evacuation (see page 1 and 2 of the final report). Mr. Ewing went ahead and provided the September 17, 2010 report "Monroe County Hurricane Evacuation Clearance Time, Final Report" and it is the DCA that will work toward agreement (via MOU) on which assumptions should be utilized for the actual final modeling and report to the Governor and Cabinet. After Dr. Ewing published his report on September 17, 2010, FDOT, Monroe County staff, and DCA staff met to review the impact of FDOT improvements constructed, under construction, or planned in the 2015 FDOT Work Plan. Aileen Boucle, AICP, from FDOT provided a letter dated October 28, 2010 (attached) and Table 1 "Roadway Configuration on US Highway 1 (Overseas Highway) and CR905/Card Sound Road in the Florida Keys, Monroe County Florida. These documents outline the outbound functional evacuation lanes for roadway segments on US1 that should be utilized in clearance time modeling, assuming all planned roadway improvements in the 2015 5 Year FDOT work plan are completed. As you can see from the WORK PLAN attached, assumptions or the modeling effort still have to be agreed upon and then after DCA completes the MOU with the local governments, DCA will run the model and they will look at the maximum build out capacity for the Florida Keys Areas of Critical State concern. Then, ultimately, the WORK PLAN calls for adjustments in the number of ROGO allocations that can be issued by the State of Florida. This in turn, may necessitate a comprehensive plan amendment or land development code amendment to our ROGO process. Please note that the State of Florida Department of Emergency Management (DEM) is simultaneously working on a regional hurricane evacuation model (which is expected to be released around December 15) which has different assumptions and different areas evacuating. 1 O O _0 a) cis c V '� Q] � o m a °' 0 0 L aD E c0 , -0 0 d(D (D `o _o CL -0 m C (i LE co V Zf sta a a) 0 �:: 0 c O N co E 3 a CO Ny O Q) O 0 CO 0 C C� .': cO S t i17, (6 a CD cc cc O V g C C CO 700 E N - 5 7' A N O G! O Q CD-(4 g �' a, ci O d L E c E E a (DN S .0.. CM N co O C ai 0 N c L 0 c O p O ca o cc co y -0 �i � `� ` y d co ID L V O. u' co �° o t£rrtirr�f��sr tit rat n � o 5 a`� c . fR fO N--� -� N cd CD 8 8 Y s w a E ,!!�� H s t I Qj LO i� C, CD r o m c0 a> E ¢ - 75 c L E > w y y > _ C ED O = = as a) o 8 Y s tt�AU tlr fti 4st ,V, O _O CL O ___ O LL 0 N E co ca 0 « t }. o ?� O CL CD N O E= Co d E E -� i N O CD 4)i ,� a c� v� o us E $ a�i o -_5 CL E N a) o j E Y 0 w 'V, d 8 m � �- O tOi .. E SE ca cm -0 O Ocri C O 4) N O O O C y N> O N as U 0 CD:A (DO - t0 CO L fO C L E O a� E O as a L a��i EEp o t�ir O Co O N d N O N U j a) a) {(tr 0 CD c d c=' .L... ccu >. a> 0a c Z -E -0 > > E > E �i o O 0 O W O O pp O d O L E CLZ U cV 0) 1 � it sad t Orftr "OS,',lt us ' d �c� m n 0 r CD - = m � N co 0 r ©+lsf�7k L Y cc L a. N CIS ))Milt i h ' w ) cmr cc a) O "W ti } � tri 1 Irt, _ '0 -� T _0 'p LL s�ll is ttt Ott y O -0 7• O 0- N d c as # vS -0 �' L .= U N— Q -0 c_ Co s t co � � c c O 2+v L c E -o w j (fir }`trif cl co c� � y £fls;tlt+;it,i w c G cL as cci c 5, E _� E E O v' 8' E a' 0 cc o �a is - co o y E = $ 0 o fn o in a o c 0 To 5 p v > aci c� co O � & M E Q� M ca = vi E U 0 lr ti' r it, O a E co E m -o a) c $ dcc VL t i7t 1t iL rrtl� Z ,,rt, E S ca E is O N is ti 0 O Ca i ca i cc$ '0 7 O E- U N (O pc N ("D c Cj N tt Ca CC 75 CO 0 j W L CO N -raO E¢ , NCD co �' it QQI 1Z-rr �irr `'' N cO a�Ni c c a' c U Y E to Y v' 7. — E�78 8 C E y= v Q :3 pia = O c o o m aci �'co ' , d c a a c� tL� Nir F f' si£,, o c w-0 � Yco m E n L�L7'u ii GDi , t 4 art 6 o c ca c a� v o c c c c� 0 a O Co 0 Q cc o (D L_ T O- 0 M in > L- N JCL F- coU 3 n d f}�ti1 rrrr�i�( t�!`6 6,Qis t' e d ai (D r r f stlx �r iV O ,'+tISiVVr,rr iriiii Sri }�,S`u rfiii c��C l( s},1i71'W,i�,r „a (D rn c 2 r� 1�0.11 r�t ti\ i & ( a y Y £'rrr`} ri�'tiJt r ti $ Irr7��i`' 1,r,r1,14 t �,11r i ti��,i 3 3��tt t}fl�i�� 1f 4chrii�i ��its ��, co y � i� yl r� yif r11 t i� r}ti(,air r;. Y ft 10 o�4>., �ffi ,hj,i�st} j C �S ti �- iS,S£(Vl\1 { `i'tisSSft2 r{rjli Vw cOi f,}fir t1 �rti�j�t 51��12}r} jrr }i��ri t'Stiif4i (Ol�J ti7ii r�l4�{{��J(�' N !y I� i�SS��fiyt�2�S tl;fylS��y�i �41i'4i;rr£,'i; f- s fl �i $ 32 E O N r �r, U�7�2f 4ir�'r1 r�,I r }tir{ �'r'14}r's ) 2��� 14tL►J,1 M i C tt,a issr�� St`}U£S'rr�j£A 1 CD > (D o Yis t� tris SG 1r1y� } ` £.f iii , Sf� rr4j�i'i�£031t i� ,11 h\$ r „ O O O \){tr off C C ri�trr 1r 1'r1t}i�r`�? fi'�1ir�; ,� titjf+ )tS�hitr � — 1 i � O O v CO 0) oa W m � E A' `A �e c> h m 'O "�6 cc m O O Y cC can N O C^{ O � V �' (�% G �.t d � V o ^i o w o 0. .�-. W 9 ,�27 o 1itS'rzniS{' ri+` Sf4it'rt� cC c N r- (Dn O N E Q c — o CO t� o cccS O N m L c1_. � aQi W � c 1 t rr(lt 'if11 �if«2,n�� T • U) cq '�_ E 7 N ' C1 C O Y co d W 3 ;7f�i1 ir�,�1t ffl < V N C .0 E 7 E N to I— c� "C 'C )� r i i� iy ltiii 1� 1 O > cc$ O cII c� ca E CIS j H C cu N E O E N 0 Z7 r pit � O c� `� 0 '6 ¢ d 0 y a) C vvri dti t r (D �p�S 1 E . .l{{ i.Tr,.7,liSG,,y L`a fq N 2 t0 O C (y� O m O C _ CCcD d N7 � Q E 5 ." 4 N d N �!+ C h d "�/J f1� O C C fE (a m a� C p) .�0� � = cO W p O 2 ?DV� d o £)sl��''+r+w LJi{yt"i�QAl itr`' N V Q 3 � d �N N T O . co W O �C 0-_ C 7+ Syyr to �^iJ�. _ U R Q d N N r 0 rA O ">p �e <C ty�r'}�\,'"'", i' iC L E d C O 0 d 1} t to t 1 i ) rr `141... raft, 1 tiM1 rr ��ts ,A��tt ,ti N t C 1 Zff })�sSti�ti {fr Ott fr i 1 niI44lft�, st tt 4 ,, t1 y{tftt rzt t�rr I r '5 lj;i j'tt(j\ eft il�ttjt�t�ll t\)-'l i , t24 �z95 �'t d St tt� i E it �Ydtl\s o a, O _ a cc C } o O o o CCD t O a •> o CC 5 C CO CM N ._ Q) F- E CC 4D cO E a E c a i7„)asp t gEwj CO N Rii"d cm C i� f rf t �. �� f6 N �5 fc$ +-+ d _ c ii{ �1il31111, )} \.CC.Q tv, �* C ,t ry {,���, N 11 V CL W N CO cO Al as 13 �i�r�tt �ii Jttl4 ,S< 'p 'C O 2 Ea> November 8, 2010 THE LJ P°d B V F' Ft , 1'ry x UTAH Department of City & Metropolitan. Planning f'i 5 534 k Ri`A 2M AAC rkr 391 8255 iAX (801 185 82:?' Monroe County Hurricane Evacuation Clearance Time — Final Report Reid Ewing, Ph.D. Professor of City and Metropolitan Planning University of Utah 1. Hurricane Evacuation Modeling Generally The federal government, under FEMA, mandates that all states have comprehensive emergency operations plans for such disasters as hurricanes. The majority of states have a two -tiered approach to emergency planning and response. Evacuation planning, response, and recovery activities are done at the local level (either county or city) while the state is responsible for coordinating local emergency management activities and state - level law enforcement and transportation. The state emergency management agency in Florida plays a larger role in managing and developing evacuation plans than other states since the state of Florida is highly susceptible to hurricanes. Evacuation models are used to estimate clearance time. Clearance time is the total time it will take to evacuate all anticipated evacuees from the vulnerable area following an evacuation order. Clearance time is calculated by adding the amount of time it takes residents of an area to prepare for an evacuation (mobilization response time) and the amount of time it takes them to leave the area (evacuation time). Hurricane evacuation clearance times are used as emergency management tools throughout the state of Florida. However, in Monroe County only, estimated hurricane evacuation clearance times are also used for regulatory and growth management purposes. Specifically, since 1992, Monroe County has used clearance times to control the rate of growth in the county, with State of Florida oversight. In 2005, the Monroe County Year 2010 Comprehensive Plan was amended to establish a three-phase evacuation process, as follows: Policv 216.1.8 In the event of a pending major hurricane (categoi�v 3-5) Monroe Comanty shall�implement the folloiring staged phased evacuation procedures to achieve and maintain an overall 24-hour hurricane evacuation clearance time for the resident population. November 8, 2010 1. Approximately 48 hours in advance of tropical storm hinds, a mandatory evacuation of non-residents, visitors, recreational vehicles (RV's), travel trailers, live- aboards (transient and non -transient), and military personnel from the Kevs shall be initiated. State parks and campgrounds should be closed at this time or sooner and entry into the Florida Kevs by non-residents should be strictly limited. 2. Approximately 36 hours in advance of tropical storm hinds, a mandatory evacuation of mobile home residents, special needs residents, and hospital and nursing home patients from the Kevs shall be initiated 3. Approximately 30 hours in advance of tropical storm hinds, a mandatory phased evacuation of permanent residents by evacuation zone (described below) shall be initiated Existing evacuation zones are as follows: I-6) a) Zone I Kev West, ,Stock Island and Kev Haven to Boca Chica Bridge (I Eff b) Zone 2 Boca (hica Bridge to West end of 7-mile Bridge (I Eff 6- l0) c) Zone 3 West end of 7-Mile Bridge to West end of Long Boat Kev Bridge (MM 10-63) d) Zone 4 West end of Long Boat Kev Bridge to CR 905 and (R 905A intersection (MM 63-106. 5) e) Zone 905A to, and including Ocean Reef (I Eff 106.E 126.5) The actual sequence of the evacuation bh y zones ill vary depending on the individual storm.. The concepts embodied in this staged evacuation procedures should be embodied in the appropriate County operational Emergency Manager??ent Plans. The evacuation plan shall be monitored and updated on an annual basis to reflect increases, decreases and or shifts in population; particularly the resident and non- residentpopulations. [9J-5.012(3)(c) lJ Objective 101.2 of the Comprehensive Plan requires Monroe County to reduce hurricane clearance time to 24 hours by 2010. The Miller Model, developed specifically to estimate clearance time for the Florida Keys, has yet to be tested with a phased evacuation scenario to see if Monroe County meets this objective. Our charge is to conduct such a test, while updating the model based on 2000 U.S. Census data, recent building permit data, the best available tourist data, all available hurricane survey results, realistic roadway link capacities, and other data that have become available since the last test. This report estimates clearance time under three- phase evacuation for a worst case Category 5 hurricane. Clearly, estimated clearance time will vary with the assumptions made in the Miller Model update. The matrix in the Appendix at the end of this report sets forth the assumptions proposed by different agencies. This update is based on the assumptions in the Ewing column, which the author views as most realistic. 2 November 8, 2010 Conventional Evacuation Models Conventional hurricane models make use of traditional urban transportation models, the same models used in long-range transportation planning. There are more than 30 transportation modeling tools that have been used for evacuation modeling. In addition, there are also several specialized transportation planning models that were developed specifically for hurricane evacuation events, including ETIS, HEADSUP, and HURREVAC. These three models are described in more detail below. There are three basic ways to model a traffic network: macro, micro and meso. The three models differ in terms of scale (geographic area) and the level of detail (how precise the analysis is). Therefore, "[u]nderstanding the potential of transportation modeling to support decision -making for evacuations hinges on identifying those decisions in the process that best lend themselves to the strengths of a particular modeling approach."' Macro models are able to represent a large geographic area such as an entire metropolitan area; however, these models cannot represent individual vehicles or people on the road network. A sub -category of macro models that are time sensitive, real-time decision support tools, are becoming increasingly popular. Micro models represent only a portion of a road such as milemarkers along an interstate. These models are helpful in modeling smaller sections of a network such as a specific roadway corridor and are able to calculate precise results since individual vehicles are tracked on the network for a small segment of time (normally 1/loth of a second). A third type of model, meso models, are able to represent larger geographic areas than micro models and at the same time are able to allow for more precise results than macro models. In addition, these models are able to represent individual roadway links and vehicles on a network; however, they are not able to represent individual lanes on each roadway segment. HURREVAC is a macro model designed by the U.S. Army Corps of Engineers for FEMA to assess hurricane evacuation scenarios. The model estimates the amount of time it will take to evacuate an area and can be used to determine the best time to begin an evacuation. The model uses information from the National Hurricane Center, flood estimates from the SLOSH model, and information on the utility of all shelters in the area. PBS&J developed the ETIS model following Hurricane Floyd. This is a macro -level modeling and analysis system which is primarily comprised of an Internet travel demand forecasting system. The system is able to predict congestion from evacuation traffic as well as traffic flows between states. It allows emergency officials to input the category of storm, the estimated participation rate, tourist occupancy rate, and destination percentages for the counties of concern. With such data, the model is able to output the level of congestion on major highways as well as tables of anticipated vehicle volumes. t Hardy, Matthrew and Wunderlich, Karl. (2007). Evacuation Management Operations (EMO) Modeling Assessment: Transportation Modeling Inventory. Pg. 19. November 8, 2010 The Florida HEADSUP program is used to manage traffic proactively during an evacuation. Although HEADSUP uses the same information as ETIS, the program is more detailed and complete. The program is able to automatically process real-time traffic data from 27 strategically located traffic counters throughout Florida in order to analyze evacuation conditions and assist in emergency management decisions. The program is also able to run hourly dynamic travel demand forecasts, impact analyses of contraflow lanes, socio-economic statistics on evacuees, a map -based user interface, a traffic model that gradually loads evacuees onto the roadway network, and an archival capability which records when key events occurred during a hurricane evacuation. The Florida Keys Hurricane Evacuation Model, widely known as the Miller Model, is a deterministic model that supplies a specific model output — clearance time — based on such inputs as the number of dwelling units and capacity of roadway links. Miller Consulting developed this hurricane evacuation model in 2000 to measure and analyze the unique characteristics of the Florida Keys and to determine the clearance time required to evacuate the Florida Keys up to Florida City, based on existing US 1 conditions. The Miller Model was designed to model the behavior of residents and tourists in responding to a mandatory hurricane evacuation order in the Florida Keys and is able to test various scenarios in order to determine the clearance time for each scenario. State -of -the -Art Evacuation Models Traditional urban transportation models are static. They do not take into account the dynamic changes that occur in travel behavior during the evacuation process. The static models assume stable conditions both in demand variables and traffic flows. Haoqiang Fu and Chester Wilmot have developed a sequential logit dynamic travel demand model for hurricane evacuation. The model considers the evacuation order as a time -dependent variable rather than a static variable and thereby analyzes both the impact of the type and timing of evacuation orders. The model divides evacuation time into discrete intervals; the probability of a household evacuating in a particular interval is the product of the probability of evacuating in that time period and the product of the probability of not evacuating in all earlier time intervals. The model is also designed to test phased evacuation. Fu and Wilmot used a small dataset from Southeast Louisiana from Hurricane Andrew to develop their dynamic model. Due to the limitations with the size of this dataset, Fu and Wilmot then estimated a similar sequential logit model using a larger dataset from South Carolina collected after Hurricane Floyd. This model is considered state-of-the-art because it is able to analyze the impact of the type and timing of evacuation orders. Fu and Wilmot used the model to better understand household evacuation behavior under different evacuation order conditions. The model can also be used to study the impact of a variety of factors such as the type and location of the residence, and storm -specific characteristics such as wind speed, forward speed, and the path of the hurricane. 4 November 8, 2010 Monroe County could benefit from developing a dynamic model for future hurricane evacuation updates. It would provide a more accurate measure of clearance time than the currently used evacuation response curves. 2. The 2001 Study While other modeling options exist and may be pursued in the future, time and budget limitations under our contract led to a decision to update a conventional model developed in the 2001 Kevs Hurricane Evacuation Vudv (the 2001 Study). The conventional model is widely referred to as the Miller Model. The model is a spreadsheet -based program executed in Microsoft Excel. The model is comprised of 39 Excel spreadsheets, 31 of which relate to individual roadway segments. The 31 roadway segments are defined by roadway cross-section, capacity, and mile markers. The model is deterministic, predicting evacuation movement link -by -link, in 2-minute increments, assuming a 30 mph average driving speed. Clearance Time There are different definitions of clearance time, depending on the hurricane model that is utilized. The 2001 Study definition is: "...the time required to clear the roadways of all vehicles evacuating in response to a hurricane situation. Clearance time begins when the first evacuating vehicle enters the road network and ends when the last evacuating vehicle reaches its destination." This definition had to be modified to account for the phasing of evacuation and the tendency of some residents to evacuate spontaneously before an evacuation order is issued. "Clearance time" begins 36 hours prior to tropical force winds when mobile home residents are ordered to evacuate (at the beginning of Phase 2), and it ends when the last evacuating vehicle exits, or passes by the northbound entrance to Florida's Turnpike on US 1 in Florida City. For purposes of determining total time to safety for evacuating vehicles, the 2001 Study added Dade County travel time to Monroe County clearance time to reflect an approximate time to get from Florida City to the evacuation shelter at Florida International University (FIU). This additional time was assumed to be 30 minutes for Category 1-2 hurricanes, and 52 minutes for Category 3-5 hurricanes reflecting addition congestion under the worst case. As we are only interested in time to evacuate to Florida City, this update does not include this additional travel time. Zone Structure When the 2001 Study was in process, a decision was made to delineate seven evacuation zones, as that was what the Monroe County's Emergency Management Division was using at the time. The Monroe County's Emergency Management Division has since transitioned to five hurricane evacuation zones. Moreover, the South Florida Regional Planning Council has opted to base the zone stricture of its evacuation model on census geography, which simplifies model updates. November 8, 2010 For this application, we held to the seven -zone stricture of the 2001 Study. The seven zones are defined by mile makers: Table 1. Mile Marker Limits for each Evacuation Zone Evacuation Zone Mile Marker Lower Keys 1 0-13 2 13-46 Middle Keys 3 46-64 Upper Keys 4 64-84 5 84-95 6 95-113 7 106-ICWW To update inputs to the Miller Model based on the 2000 Census, it was necessary to determine how census geography relates to the seven 2001 Study evacuation zones. We used a combination of maps provided in the 2001 Kevs Hurricane Evacuation ,Study and descriptions of the zonal boundaries to produce the following correspondence table (Table 2). Table 2. Zone Stricture for Updated Miller Model (2008) Traci B,oc1 Arou' ,1?ertage crf Bock �rc�ul3 ui'ZQnc ' Zone 1 (Key West to Saddle Bunch Channel Bridge - mm 0-13) 9726 All block groups 100% 9725 All block groups 100% 9724 All block groups 100% 9723 All block groups 100% 9722 All block groups 100% 9721 All block groups 100% 9720 All block groups 100% 9719 All block groups 100% 9718 All block groups 100% 9717 All block groups 100% Zone 2 (Saddle Bunch Bridge to Knight Key Channel - mm 13-46) 9716 All block groups 100% 9715 All block groups 100% 9714 All block groups 100% Zone 3 (Knight Kev 9713 All block groups 100% 6 November 8, 2010 C,esus praot Block Crroii�a;' 'erQentage Q1ock roup in':one ' Channel to Long Key Viaduct - mm 46-64) 9712 All block groups 100% 9711 All block groups 100% 9710 2 100% 9710 3 100% Zone 4 (Long Kev Viaduct to Whale Harbor Channel - mm 64-84) 9710 1 100% 9709 1 40% 9709 2 45% 9709 3 100% 9709 4 100% 9709 5 100% Zone 5 (Whale Harbor Channel to Milemarker 95 - mm 84-95) 9709 1 60% 9709 2 55% 9708 All block groups 100% 9707 All block groups 100% 9706 3 100% Zone 6 (along U.S. 1 - mm 95-113) 9706 1 100% 9706 2 100% 9705 All block groups 100% 9704 All block groups 100% 9703 All block groups 100% 9702 1 40% 9702 3 60% Zone 7 (along CR 905 - mm 106-ICWW) 9702 1 60% 9702 2 100% 9702 3 40% 9701 All block groups 100% November 8, 2010 Inputs The Miller Model requires the following inputs related to housing, evacuee behaviors, and road network performance. • How many dwelling and tourist units exist in the evacuation area; • What fraction of the dwelling and tourist units will be occupied at the time of evacuation; • How many people will leave their dwellings to go someplace safer (i.e., evacuation rate or evacuation participation rate); • When evacuees will leave, with respect to when evacuation orders are issued; • What effect a policy of phased evacuation will have; • Where the evacuees will go, in terms of ultimate destinations inside or outside the county; • How many vehicles will be used in the evacuation; • Where evacuating traffic will load onto the road network; • How much background traffic will be using the road network at the same time; • How much traffic can be handled by critical links in the road network; The following chapter outlines sources of data, methods of estimation, and values for each of the above used in our update of the 2001 Study. 3. Update of the 2001 Study Numbers of Dwellings and Tourist Units 2001 Study Evacuating population comes from three types of units: 1) permanent dwelling units, 2) mobile home units, and 3) tourist units. The 2001 Study began with the official number of dwelling units as of 1990 from the U.S. Census. Monroe County Planning Department then provided numbers of new units based on certificates of occupancy (CO) issued each year. The number of COs was summed, cumulatively, from 1990 to 1999. After 1999, the methodology followed by the County shifted to the potential number of dwelling units available under the permitting guidelines of the Rate of Growth Ordinance (ROGO). 0 November 8, 2010 Update The number of permanent dwelling units and mobile homes was determined from the 2000 U.S. Census, updated to reflect new dwellings occupied between 2000 and 2008 (see Tables 3 and 4). Permanent dwellings in 2000 included all census categories of permanent strictures from single-family detached to multifamily with 50 or more units. Mobile homes included census categories of "mobile home" and "RV, boat, van, etc." The decision to include the latter with the former was prompted by belief that permanent residents living in RVs (many in mobile home parks), boats, vans, etc. would behave more like mobile home residents than tourists in an evacuation. Permit data for new residential units issued from 2000 through 2008 were provided by the Monroe County Building Department and the equivalent departments of the five incorporated cities in Monroe County —Key West, Islamorada, Key Colony Beach, Layton, and Marathon. Post-2000 unit counts were added to 2000 unit counts to obtain current estimates of dwelling units by evacuation zone. Tourist unit data was collected from the Department of Profession and Business Regulation. This department licenses hotels, motels, bed and breakfasts, timeshares and vacation rental units — all of which were included in the update. The data from DPBR were geocoded by Bryan Davisson, the GIS Planner in Monroe County's Growth Management Department. Table 3. Permanent Dwelling Units in 2000, constricted and occupied between 2000-08, and total in 2008 Zone 2000 2000-08 Key West 2000-08 Islamorada 2000-08 Marathon 2000-08 Key Colony Beach 2000-08 Layton 2000-08 County 2008 Total 1 14,509 319 280 15,108 2 6,143 360 6,503 3 6,972 124 170 47 7,313 4 1,880 21 3 1,904 5 5,095 169 42 5,306 6 5,093 242 5,335 7 1,310 0 1,310 Total 41,002 319 169 124 170 21 974 42,779 9 November 8, 2010 Table 4. Mobile Home Units in 2000, permitted between 2000 and 2008, and in 2008 Zone 2000 2000-08 2008 1 2,496 2,496 2 1,751 1,751 3 1,940 1,940 4 720 2 722 5 1,219 1 1,220 6 2,459 1 2,460 7 8 8 10,593 4 10,597 Table 5. Tourist Units in 2008 Zone 2008 lodging 2008 vacation rental 2008 timeshare 2008 Total 1 8,148 0 0 8,148 2 491 23 0 514 3 2,997 29 19 3,045 4 1,734 2 1 1,737 5 576 0 0 576 6 1,960 3 14 1,977 7 36 0 19 55 15,942 57 53 16,052 Occupancy Rates 2001 Study The Project Steering Committee (PSC) identified "% Occupancy of Dwelling Units" as a critical variable. The PSC used 1990 Census data to determine the occupancy rates during the month of April (when the Census data are collected). For tourists, the occupancy rate utilized was from the 1991 Hurricane Evacuation Analysis of the Monroe County Comprehensive Plan and the 1995 update, both prepared by PBS&J. The occupancy was estimated as 45% on the low end and 75% on the upper end. The Project Steering Committee studied these numbers and decided to estimate the occupancy rate by subregion of the Keys. Actual rates, based on specific knowledge of the Project Steering Committee members, were used whenever available. For example, 10 November 8, 2010 an occupancy rate of 72% was used for Key West since members knew that overall occupancy rate here was higher than the rest of the county. Update Occupancy rates for permanent dwellings were determined by zone from the 2000 Census (see Table 6). Occupancy rates for the county as a whole appear to have declined by about 20 percent between the 2000 Census and the 2008 American Community Survey. We therefore produced a second set of occupancy rates, prorating 2000 occupancy rates by zone to account for this decline (see Table 6). Table 6. Occupancy Rates for Permanent Dwellings and Mobile Homes (2000 and 2008 estimate) Zone Percent Occupied Housing Units — 2000 Census Percent Occupied Housing Units — Adjusted for 2008 American Community Survey 1 84% 67% 2 67% 54% 3 59% 47% 4 44% 35% > 58% 46% 6 65% 52% 7 34% 27% To update tourist occupancy rates, we referred to Smith Travel Research's latest Trend Report, submitted annually to Monroe County's Tourist Development Council. Occupancy rates have remained relatively constant over the years. During the hurricane season (June 1 through November 30), July is the highest occupancy month, while September is the lowest. We used July 2008 values (see Table 7). This is a worst -case assumption, since the peak of Atlantic hurricane activity is in September, the month with the lowest occupancy. 11 November 8, 2010 Table 7. Occupancy Rates for Tourist Units (July 2008) Zone Percentage Occupied Units 1 (Kev West) 82% 2 71% 3 71% 4 71% 5 71% 6 (Keys Largo) 77% 7 71% Evacuation Participation Rates 2001 Study To estimate evacuation participation rates, the 2001 Study relied heavily on a survey conducted by Dr. Carnot Nelson in 1989. The assumed evacuation participation rates are shown in Tables 7 and 8. All are taken from Dr. Nelson's behavioral analysis, except participation rates for tourist units which were assumed to be 100%. Dr. Nelson had suggested lower numbers. Nelson's survey was done before Hurricane Andrew, and it simply asked people what they intended to do in response to a number of hypothetical hurricane threats. Intended - response data may be unreliable predictors of actual evacuation behavior. Much more information has become available since Nelson's pre -Andrew survey (Baker 2000): • A University of Florida group conducted a survey following Andrew, not only asking what people did in Andrew, but also using the very same intended - response questions previously used by Nelson. • James Mattson conducted a survey following Andrew, dealing with Andrew response and intended response in future storms. • Dr. Earl Baker did a survey following Andrew for the National Science Foundation that documented response in Andrew, perceptions of vulnerability, confidence in constriction, and intended responses in future threats. • Following Georges, FIU conducted a survey documenting response to Georges as well as asking about certain subjects that could have a bearing on future response. • Also following Georges, the Monroe County School Board had public school students take home a questionnaire asking what their households did in Georges. • Dr. Earl Baker conducted interviews in the Lower Keys as part of a post -Georges survey for the Corps of Engineers and FEMA. It dealt with response to Georges 12 November 8, 2010 as well as vulnerability perception, concerns about traffic congestion, and future response. Dr. Earl Baker conducted an additional survey in the Lower Keys, dealing with response to Georges but also posing several hypothetical threat scenarios and evaluating the effect on intended response of roadway improvements and having refuges of last resort in Key West. Following Hurricane Ivan, a Post -Ivan Behavioral Analysis was prepared for the Federal Emergency Management Agency and the U.S. Army Corps of Engineers in September 2005. A total of 200 interviews were conducted in Monroe County. The questionnaire asked questions regarding evacuation decisions and behavior, home mitigation and/or preparation, household circumstances, economic impacts, and household information needs. The South Florida Behavioral Survey was conducted in 2007-2008 as part of Statewide Regional Evacuation Study Program. The primary aim of the survey was to provide data to assist in deriving evacuation behavioral assumptions for transportation and shelter analyses. In each non -coastal county of the state 150 interviews were conducted randomly by telephone. In each coastal county of the state, 400 interviews were conducted. Baker Study Based on actual and intended responses to hurricanes, from several surveys after Hurricanes Georges, Andrew, and Irene, Professor Earl Baker at Florida State University derived most probable evacuation participation rates for a number of hurricane threat scenarios. Earl "Jay" Baker is an associate professor of geography and an expert in the field of hurricane evacuation. His research is focused on how people respond to warning and evacuation orders and how emergency managers are able to use forecasts to implement evacuation plans. He has studied peoples' vulnerability perceptions and hurricane preparedness in most areas of the Gulf of Mexico and Atlantic coasts. Table 8 provides Baker's best estimates of participation rates for Category 5 storms approaching the Keys from the south, posing a greater risk to the Lower Keys. Table 8 also provides his best estimates of participation rates for storms at latitudes similar to Andrew, posing a greater risk to the Upper Keys. The table assumes mandatory evacuation orders and aggressive actions by public officials to educate the public about appropriate responses. 13 November 8, 2010 Table 8. Evacuation participation rate assumptions for Category 5 hurricanes approaching from different latitudes, aggressive mandatory evacuation ordered and improved public education regarding vulnerability (Baker 2000) from latitudes south of Key West from latitudes similar to Andrew Lower Keys 90 35 Middle Kees 95 95 Upper Keys 95 100 South Florida Behavioral Survey The 2008 South Florida Behavioral Survey asked whether respondents intended to evacuate their homes for some place safer if mandatory evacuation notices were issued due to potential flooding (see Table 9). The question was asked for both Category 3 and 5 hurricanes. Results weren't presented for Category 4 hurricanes. The Category 5 results are most relevant to this worst -case analysis. Table 9. Would Leave Home if Mandatory Evacuation Notice is Given for a Category 5 Hurricane N Yes No Don't know/depends Yes plus Don't know/depends Monroe 400 88% 8% 4% 92% Key' West 100 89% 9% 3% 92% Lower Kees 100 91% 6% 3% 94% Middle Keys 100 90% 7% 3% 93% Upper Keys 100 84% 8% 8% 92% Perhaps a better predictor of evacuation participation than intended response to hurricanes is perceived vulnerability to both wind and water in hurricanes of different intensities. Table 10 reports Monroe County responses to the question of whether respondents would remain safe in a Category 4 hurricane (Category 5 results weren't released). 14 November 8, 2010 Table 10. Safe from Wind and Water in a Category 4 Hurricane N Yes No Don't know/depends Monroe 400 15% 80% 5% Key West 100 19% 76% 4% Lower Kees 100 11% 81% 7% Middle Keys 100 15% 83% 1% Upper Keys 100 13% 79% 8% Monroe County residents were also asked if they left home during Hurricanes Georges (a Category 2), Ivan (a tropical depression as it approached Florida), and Wilma (a Category 2 hurricane in Monroe County). Hurricane Georges prompted 38% of households in the Monroe County region to evacuate, with the Middle Keys reporting the highest participation (50%). Hurricane Ivan caused 28% of households in Monroe County region to evacuate, with the Upper Keys reporting the highest participation (34%). Hurricane Wilma caused 32% of households in Monroe County to evacuate, with the Lower Keys reporting the highest participation (37%). These results are for low -intensity hurricanes; no Category 4-5 hurricanes have hit the Keys in recent years. Update The worst case is a Category 5 hurricane that approaches from latitudes below Key West, with aggressive mandatory evacuation ordered and improved public education regarding vulnerability (see Table 11). Baker suggests that 90-95% of residents might evacuate under such circumstances. While no clear geographic pattern of evacuation compliance emerges from the various surveys, we will go an upper bound evacuation participation rate equal to Baker's recommended rates. In this worse case, a 100% evacuation rate will be assumed for mobile home and tourist units. Actual evacuation rates during past hurricanes have reportedly been much lower than this worst case. True, these were less intense hurricanes than posited here, but it seems likely that respondents overstate their willingness to evacuate when asked to speculate in surveys. We will therefore conduct a sensitivity test of clearance time, assuming a lower bound evacuation participation rate of 70-75% for permanent dwellings in response to a more typical hurricane. Table 11. Category 5 Storm Evacuation Participation Rates Mobile Homes Tourist Units Other Units Lower Keys (Zones 1 & 2) 100% 100% 70-90% Middle Keys (Zone 3) 100% 100% 75-95% Upper Keys (Zones 4, 5, 6 & 7) 100% 100% 75-95% 15 November 8, 2010 Evacuation Timing Evacuation timing refers to when evacuees depart their residences. While some spontaneous evacuation occurs, it is unusual for more than 15% of the eventual evacuees to have departed before officials issue evacuation orders. Departures then occur depending upon the urgency perceived by evacuees. 2001 Study The 2001 Study uses tables to represent the rate at which evacuating traffic enters U.S. 1. The exact number of hours over which the traffic is loaded is not terribly important. The main thing is that the scenarios reflect a range of plausible response distributions, based on the timing of evacuation orders prior to landfall, to assess the sensitivity of clearance times to those variations. The 2001 response curves don't reflect the fact that some evacuees will leave before an evacuation order is issued. That is clearly wrong. Dr. Balser calls 10% spontaneous evacuation a conservative figure. Baker Study Dr. Baker developed the curves in Figure 1. They indicate how promptly evacuees depart when evacuation orders are issued under three scenarios of urgency. "Late, normal, and early" refer to when evacuation orders were issued relative to expected arrival of a hurricane. These curves assume 10% spontaneous evacuation even before the evacuation order is issued. Figure 1. Early, normal, and late evacuation timing curves Typical Evacuation Timing Cur 100 (n a> C= 80 m w �60 c U N 0-40 m E20 U M 6 4 2 0 2 4 6 8 10 12 14 16 18 20 22 24 Hours Before/After Evacuation Order — Late m Normal Early 16 November 8, 2010 Based on evacuation response to Hurricanes George and Andrew, Baker developed the two-day curve in Figure 2. This response curve accounts for early evacuees even before evacuation orders are issued. At least for strong hurricanes, Baker concluded that such a curve could apply to Monroe County. Figure 2. Two-day evacuation timing response curve Un100 U > 80 W O C 60 U N d > 40 m E U 20 Update Two -Day Evacuation Curve 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 Hours After Evacuation Notice The three Baker curves in Figure 1 seem most applicable to evacuation scenarios for Monroe County, where a mandatory evacuation order is issued early, at a normal time, or late. The fact that Baker provides three different curves allows us to perform sensitivity tests on evacuation timing assumptions. One anomaly associated with the Baker curves is that the clearance time cannot be less than 24 hours when an evacuation order is issued early, which is arguably the scenario which involves the least risk to the public. Therefore, in assessing clearance time, primary emphasis will be placed on the late response scenario. 17 November 8, 2010 Effect of Phased Evacuation 2001 Study In the 2001 Study, all residents and tourists were assumed evacuate at the same time. Update In 2005, Monroe County adopted a mandatory phased evacuation policy as part of the update of its comprehensive plan. This phased evacuation requires that all tourists, recreational vehicles, military and live aboard vessels begin to evacuate from the county 48 hours in advance of tropical force winds. Next, mobile homes and special needs residents will receive the order to evacuate 36 hours in advance of tropical force winds. Last, the residents living in permanent dwelling units will receive the order to leave 30 hours in advance of these winds. The Miller Model had not been used to test phased evacuations before and therefore needed to be adapted. This was done by having separate response curves and trip tables for mobile home residents and permanent dwelling unit residents, with a six hour lag between the former and the latter. The two groups of evacuees are added together where their response curves and trip tables overlapped. The Miller Model had to be significantly modified to represent a phased evacuation. Both groups of residents were assumed to evacuate according to Dr. Baker's late response curve in Figure 1, with overlap between the two groups starting at 30 hours prior to tropical force winds. Essentially, since the late response curves show evacuees leaving home over approximately a 12 hour period, there is six hours of overlap in departures between the groups. Of course, after that, they are on the road together for the remainder of the evacuation trip. Handling tourist evacuees involved a judgment call. Under phased evacuation, the tourist evacuation order will be issued 48 hours before tropical force winds, or 12 hours before the evacuation order for mobile home residents. Dr. Baker's most recent report, based on 2009 surveys of hotels, motels, resorts, bed and breakfasts, seasonal housing rentals, and recreational vehicle parks, suggests that 30 percent of tourists evacuate spontaneously before the order is issued, and another 40 percent of tourists evacuate in the first 12 hours after the order (see Figure 3). This leaves 30 percent of tourists to evacuate at the same time as the mobile home park residents. To simplify the model calculations, this 30 percent of tourists was simply added to the mobile home park total and assumed to evacuate following the same response curve. Figure 3. Tourist Evacuation Timing 18 November 8, 2010 Keys Tourist Evacuation Timing Untnreighted and Weighted by Number of Units and Vehicles 100 Ar 90 80 70 60 so Q ,fix WC^ighttd by Uilit^; y 40 z , Weighted by Car z 30 -- ----®..._F E v 20 10 0 C G 12 18 24 30 36 42 48 Flours Following, Evacuation Order Source: Earl J. Baker, Behavioral Assumptions for Hurricane Evacuation Planning in Monroe County, prepared for the Department of Community Affairs, September 2009, p. 4. Destinations 2001 Study Based on Dr. Nelson's research, the 2001 Study had four possible destinations for the resident evacuees: 1) Monroe County public shelter, 2) Monroe County motel, 3) Monroe County friend or relative, and 4) Out of Monroe County. Baker Study Based on several surveys of actual and intended behavior after Hurricanes Georges and Andrew, the Baker 2000 report indicates the most likely percentage of evacuees from the three different areas of the Keys who will go to destinations outside of Monroe County for different categories of storm intensity (see Table 12). Table 12. Planning assumptions for percent of evacuees leaving Monroe County, aggressive mandatory evacuation ordered throughout Monroe County for all categories Cat 3-4 Cat 5 Lower Keys 80 90 Middle Keys 90 95 Upper Keys 95 100 19 November 8, 2010 South Florida Behavioral Survey The 2008 survey asked respondents where they would go if they evacuated for hurricanes of different intensities. Results for Category 5 hurricanes are shown in Table 13. Table 13. Evacuation Destination (Category 5) N Own neighborhood Own county Someplace else in Florida Someplace outside Florida Don't know . TVIonxoe` 3fl4" 3%0 7°l0«6%' 11 11 P4�: Key' West 72 7% 13% 52% 14% Lower Keys 79 2% 7% 69% 19% 3% Middle Keys 77 1% 1% 71% 21% 6% Upper Keys 76 2% 6% 68% 15% 8% Data are available on the destinations of evacuees during three previous hurricanes (Tables 14-16). The great majority of evacuees leave the county. Residents of Key West are most likely to leave the county, while residents of the Upper Keys are least likely to leave the county (though a majority still do). Table 14. Destinations of Evacuees (Hurricane Georges) N Own neighborhood Own county Someplace else in Florida Someplace outside Florida Don't know . 1VIoTiTof $� '%o1% 70'0o tQ/o Key' West 20 2% 5% 91% 1% 0% Lower Keys 18 0% 2% 68% 25% 5% Middle Keys 26 1% 19% 79% 1% 0% Upper Keys 16 8% 37% 46% 8% 0% 20 November 8, 2010 Table 15. Destinations of Evacuees (Hurricane Ivan) N Own neighborhood Own county Someplace else in Florida Someplace outside Florida Don't know . 1VIoTiTof76Q/o72%='.o 11 Key' West 22 % 3% 93% 4% 0% Lower Keys 25 5% 1% 75% 9% 10% Middle Keys 17 0% 8% 79% 12% 0% Upper Keys 20 0% 24% 56% 20% 0% Table 16. Destinations of Evacuees (Hurricane Wilma) N Own neighborhood Own county Someplace else in Florida Someplace outside Florida Don't know Monrt2e 82 1`0fn 11°ln 81% 5°ln77, Q°la Key West 20 4% 4% 91 % 1 % 0% Lower Keys 27 0% 3% 84% 11% 2% Middle Kevs 13 0% 11% 89% 0% 0% Upper Keys 22 0% 30% 62% 8% 0% Update The survey data indicate that the majority of evacuees from Monroe County would leave the county and evacuate to another county within the state of Florida. Beyond this generalization, the data are difficult to interpret. The intended response and actual response questions point in different directions, with the percentages intending to leave the county increasing as you move north from the Lower Keys to Middle Keys to Upper Keys. But the percentages actually leaving during past hurricanes decrease as you move north. Most likely the small numbers of evacuees during past hurricanes are atypical of the larger populations. We will assume that 90% of evacuating residents from Lower Keys (Zones 1 and 2) will leave the county, that 95% of evacuating residents from the Middle Keys (Zone 3) will leave the county, and that 100% of evacuating residents from the Upper Keys (Zones 4 through 7) will leave the county. These assumptions are in line with Dr. Baker's recommendations and the original Miller model. 100% of tourists are assumed to leave the county. 21 November 8, 2010 Vehicle Use Not all vehicles available to households are used in evacuations. Vehicle use is predicted well by hypothetical response data. 2001 Study The source of the vehicle usage rates used in the 2001 Study is not specified. It was assumed that 69 to 71% of available vehicles would be used. Baker Study Dr. Balser states that the normal range for vehicle usage is 65% to 75%. Based on behavior during Hurricane Georges, the Baker 2000 report recommended that for planning purposes, it be assumed that 70% of the vehicles available to evacuating households will be used, and 10% of those households will pull a camper, trailer, or boat or take a motor home. South Florida Behavioral Survey The 2008 survey asked how many vehicles would be available to a household that could be used to evacuate, and how many vehicles would a household take if they evacuated? As can be seen from Table 30, the percent of available vehicles that would be used in an evacuation varies from a low of 72% in the Lower Keys to a high of 9 1 % in Key West. Table 30. Vehicle Availability and Use During an Evacuation N Available Vehicles % of available % of vehicles used in vehicles used households evacuation in evacuation with no vehicle .Ulonroe '4a0 1.9`i.4 81%% " Kev West 100 1.5 1.5 91% 10% Lower Kees 100 2.6 1.3 72% 2% Middle Kees 100 1.8 1.3 79% 2% Upper Keys 100 1 1.8 1 1.4 80% 3% Update The South Florida survey data are the most recent, and we believe the most accurate data available. The one exception is the very high vehicle usage rate for residents of Key West, out of line with all the other data available. Baker reports that residents of Key West used 1.11 vehicles per evacuating household during Hurricane Georges. That amounts to about 80% of the vehicles owned by households in Key West. We therefore 22 November 8, 2010 assumed the following vehicle usage rates for residents: 80% vehicle usage for Key West (Zone 1); 72% vehicle usage for the rest of the Lower Keys (Zone 2); 79% vehicle usage for the Middle Keys (Zone 3); and 80% vehicle usage for the Upper Keys (Zones 4-7). We assumed 100% vehicle usage rates for tourists. Background Traffic Background traffic is the measure of vehicles using the roadways for reasons other than hurricane evacuation. The 2001 Study defines background traffic as including: out -of - County traffic (business trips and recreational trips), non -evacuating vehicles conducting hurricane preparation trips, typical day commuting trips, etc. In sum, this traffic is comprised of non -evacuating vehicles on the road. Background traffic increases the level of traffic on the roadway system and therefore, has a direct effect on clearance time. This traffic is comprised of non -evacuating traffic and includes trips to run errands and buy hurricane supplies. 2001 Study The 2001 Study used approximations of background traffic based on recorded traffic volumes. This background traffic affects processing time through each of the 31 links and, eventually, this background traffic declines as the evacuation occurs and decreases to zero background vehicles at the end of the evacuation. For example, if a 12 hour response curve is selected for modeling purposes, the background traffic is 100% of the actual recorded count at hour one of the evacuation and zero at hour 12. A uniform distribution is assumed for the rate of decline of the background traffic. Update We have no basis for refinement of the 2001 Study background traffic assumptions. Number and Capacity of Critical Links 2001 Study The Miller Model has 31 outbound evacuating links. It relies on the critical link concept. This concept means that the evacuation time is mainly affected by the link with the highest demand to service volume ratio. This link experiences the longest delay due to the overload of evacuating vehicles. This link, the critical link, is not static and can shift due to either demand changing by link or from capacity improvements to a link. A critical variable in the determination of evacuation time is the assumed capacity of roadway links. The Miller Model takes the capacity of uninterrupted flow highways (essentially freeway quality roads) and makes downward adjustments to account for driveways and intersections. There are two potential problems with this procedure. First, 23 November 8, 2010 U.S. 1 isn't an uninterrupted flow facility but rather a state signalized arterial, whose capacity is determined using different formulas. Second, the downward adjustments are essentially arbitrary as opposed to empirically based. Update The Florida Department of Transportation (FDOT) has recommended updates to the 2001 Study to reflect the addition of auxiliary lanes and evacuation shoulders. These additions include: a. Completed projects from Table 18 of the 2001 Keys Evacuation Study b. Projects under constriction from Table 18 of the 2001 Study Projects funded in the current work program from Table 18 in the 2001 Study Table 31 compares the number of functional evacuation lanes in the original Miller model to and the number in the FDOT update. There will be substantial functional capacity added to critical links by 2015. Based on the concept of "maximum sustainable evacuation traffic flow rates," FDOT has recommended a reduction the 2001 Study flow rates for several links. The recommended rates take into account site -specific capacity studies, observational studies of actual hurricane evacuations, and traffic simulation runs. The FDOT rates are the best available. Values are compared in Table 31. Table 31. Maximum Sustainable Flow Rates per Hour Link Name Milemarkers From TTO 2001 Functional Evacuation Lanes 2015 Functional Evacuation Lanes 2001 Flow Rates Per Lane Total 2010 FDOT Flow Rates Per Lane Total Al 2.0 1 4.0 1 2 2 900 1 1,800 900 1 1,800 A2 4.0 9.0 2 2 900 1,800 900 1,800 B 9.0 17.0 1 1 1 1,350 1,350 1,100 1 1,100 C 17.0 22.0 1 1 1 1,350 1,350 1,100 1 1,100 D1 22.0 24.0 1 1 1,350 1,350 1,100 1,100 D2 24.0 25.0 1 1 1 1,350 1,350 1,100 1,100 D3 25.0 30.0 1 1 1,350 1,350 1,100 1,100 E 30.0 34.0 1 2 1,050 1,050 1,050 2,100 F1 34.0 35.2 1 1 1,350 1,350 1,100 1,100 F2 35.2 36.5 2 2 1,350 2,700 1,100 2,200 F3 36.5 37.5 1 1 1,350 1,350 1,100 1,100 G 37.5 47.0 1 1 1,500 1,500 1,200 1 1,200 H1 47.0 48.0 1 2 1,350 1,350 1,100 2,200 H2 48.0 50.2 2 2 900 FMIJ 900 1,800 50.2 50.8 2 2 900 900 1,800 24 November 8, 2010 Link Name Milemarkers From TTO 2001 Functional Evacuation Lanes 2015 Functional Evacuation Lanes 2001 Flow Rates Per Lane Total 2010 FDOT Flow Rates Per Lane Total 12 50.8 54.0 2 2 900 1,800 900 1,800 A54.0 54.5 2 2 900 1,800 900 1,800 J2 54.5 58.0 1 2 1 1,350 1,350 1,100 1 2,200 K 58.0 74.0 1 2 1 1,350 1,350 1,100 2,200 L 74.0 80.0 1 2 1,350 1,350 1,100 2,200 M1 80.0 83.5 1 2 1,350 1,350 1,100 2,200 M2 83.5 85.6 1 2 1,350 1,350 1,100 2,200 N 85.6 90.0 1 2 1 1,350 1,350 1,100 2,200 O 90.0 100.0 2 3 900 1,800 900 1 2,700 P 100.0 105.0 2 3 900 1 1,800 900 1 2,700 Q 105.0 106.3 2 3 900 1,800 900 2,700 R1 106.3 126.5 1 2 1,500 1,500 1,200 2,400 R2 126.5 HEFT 2 3 900 1 1,800 900 2,700 S 106.3 Int CR 905 / CR 905 A 1 1 1 1,350 1 1,350 1,100 1 1,100 T Ocean 1 Reef 1 Int CR 905 / CR 905 A 1 1 1 1,350 1 1,350 1,100 1 1,100 U Int CR 905 / CR 905 A US 1 1 1 1,350 1,350 1,100 1,100 Additional Clearance Time to Reach Shelter Miller Model The Miller Model added a fixed 30 minutes (category 1 or 2) and fixed 52 minutes (category 3-5) to the clearance time for the trip from Florida City to the public shelter at FIU. One of the weaknesses of the Miller Model is that it assumes a fixed time for all vehicles to travel to the FIU shelter and it does not include the effects of traffic from Miami -Dade County. The South Florida Regional Planning Council was charged with creating a model to address this deficiency. However, that model is not available at the time of this writing. Updated Miller Model Following an administrative law judge's opinion, where an opposing counsel challenged the end point of evacuation, the end point for hurricane evacuation clearance time estimates is the beginning of the Florida Turnpike in Florida City. The Department of 25 November 8, 2010 Community Affairs concurs with this end point for Hurricane Evacuation Clearance Time modeling. Therefore the final clearance time estimates do not include the 30/52 minutes to travel from Florida City to FI J. Clearance Time Estimates Table 32 provides clearance times for 12 different scenarios. The 2000 occupancies are those in the first column of Table 6. They reflect occupancies at the time of the 2000 Census. The 2008 occupancies reflect a downward adjustment in occupancies county- wide according to the 2008 American Community Survey. The low participation rates are the suggested lower bound rates for permanent dwelling units in a Category 5 hurricane coming from the southeast (70-75%). The high participation rates are the suggested upper bound rates for the same scenario (90-95%). The three maximum flow assumptions are those associated with the original Miller Model (2001 lane configuration with Miller maximum flow rates), a combination of Miller and FDOT assumptions (2001 lane configuration with FDOT maximum flow rates), and the FDOT update (2015 lane configuration with FDOT maximum flow rates). Clearance time is measured from the time of the evacuation order for permanent dwelling unit residents until the last evacuating vehicle reaches Florida City. The updated Miller Model puts time zero at 36 hours before tropical force winds, when the evacuation order is issued for mobile home residents. Therefore, we subtracted six hours from the Miller Model clearance time outputs to arrive at clearance times relative to the evacuation order for permanent dwelling residents. The longest clearance times are, of course, associated with the 2001 lane configuration and the lower FDOT maximum flow rates. The shortest are associated with the 2015 lane configuration, which includes additional lanes compared to 2001, and the FDOT maximum flow rates. Clearance times associated with the 2001 lane configuration and Miller's higher flow rates are intermediate. The difference between these clearance time estimates and those in my report of September 17, 2010 are due entirely to the exclusion of travel time from Florida City to the FIU shelter in these most recent estimates. The earlier report erroneously said that a fixed 52 minutes had been added to the Miller Model's clearance time estimates to account for this last leg of the evacuation. In fact, 52 minutes were added to the clearance time for the "High Participation" scenario but only 30 minutes were added to the clearance time for the "Low Participation" scenario, in keeping with the reduced traffic volumes. My apologies for this erroneous statement. The reader will note that using a simple model like the Miller Model, based on fixed capacities and speeds on the different links, clearance time is not sensitive to the assumed participation rate because there is ample capacity to handle the additional traffic with the additional lanes constricted or planned by FDOT. The clearance time reflects unimpeded travel by the last evacuating vehicle from Key West to Florida City. 26 November 8, 2010 Table 32. Clearance Times (relative to the permanent unit evacuation order) Low Occupancies (2001) High Occupancies (2008) Occupancyby Zone 1=670'o; 2=540'o; 3=470'o; Occupancyby Zone 1 84°'o; 2=670'o; 3=590'o; 4=350o; 5=460o; 6=520'o; 7-2700 4=440'o; 5=580'o; 6=650'o; 7-3400 Low High Low High Participation Participation Participation Participation Approx 70% Approx 90- Approx 70% Approx 90- 95% 95% 2001 Lanes/2001 16 hours 16 18 hours 50 18 hours 32 22 hours 6 Miller Flow Rates minutes minutes minutes minutes 2001 Lanes/2010 18 hours 58 22 hours 28 22 hours 8 27 hours 2 FDOT Flow Rates minutes minutes minutes minutes 2015 Lanes/2010 16 hours 16 16 hours 16 16 hours 16 18 hours 40 FDOT Flow Rates minutes minutes minutes minutes 2015 Lanes/2010 16 hours 16 17 hours 16 17 hours 4 20 hours 16 FDOT Flow Rates minutes minutes minutes minutes (without outbound shoulder from mm 90 to mm 106) 27 November 8, 2010 Appendix 2000 Miller Ken Metcalf PBS&J Hurricane Model (1990 2004 Miller 2008 Statewide Regional Miller Model Reid Ewing Evacuation Analysis Census & Update Evacuation Study Program Analysis - Recommendations Dec.1991 (1990 PSC) (2000 Census) South Florida Behavioral Summary of Report Census) Final Report Survey Report 2000 Census in 2001 Same behavioral Sample size (n=400) parameters of 1989 ACOE stud 7 evac zones 7 evac 7 evac zones 7 evac zones zones Zone 1 - 2.44 Zone 1 - 2.44 Zone 1 - 2.44 Zone 1 - 2.35 2-2.31 2-2.31 2-2.31 2-2.21 3-2.25 3-2.25 3-2.25 3-2.18 4-1.97 4-1.97 4-1.97 4-2.08 5-2.27 5-2.27 5-2.27 5-2.27 Number of People per 6 - 2.27 6 - 2.27 6 - 2.27 6 - 2.27 M.H. Unit 7-2.11 7-2.11 7-2.11 7-1.74 Zone 1 - 2.44 Zone 1 - 2.44 Zone 1 - 2.44 Zone 1 - 2.35 2-2.31 2-2.31 2-2.31 2-2.21 3-2.25 3-2.25 3-2.25 3-2.18 4-1.97 4-1.97 4-1.97 4-2.08 5-2.27 5-2.27 5-2.27 5-2.27 Number of People per 6 - 2.27 6 - 2.27 6 - 2.27 6 - 2.27 Permanent Unit 7 - 2.11 7 - 2.11 7 - 2.11 7 - 1.74 Zone 1 - 2.90 Zone 1 - 2.90 Zone 1 - 2.90 Zone 1 - 2.90 2-3.76 2-3.76 2-3.76 2-3.76 3-2.75 3-2.75 3-2.75 3-2.75 4-2.53 4-2.53 4-2.53 4-2.53 5-12.80 5-3.00 5-3.00 5-3.00 6-12.90 6-3.00 6-3.00 6-3.00 Number of People per 7 -12.90 7 - 3.00 7 - 3.00 7 - 3.00 Tourist Unit Zone 1 - 1.80 1 - 1.35 1 -1.36 Key West 1.5 Key West Vehicle/occupied 1 - 1.36 2-1.80 2-1.76 2-1.74 1.5 unit 2-1.73 3 - 1.82 Zone 1 -1.36 4-2.00 2-1.73 5 - 2.00 3 -1.39 3 -1.56 Lower 2.6 Lower 1.3 3 -1.56 3 -1.60 Number of Vehicles per 6 - 2.00 4 -1.63 Unit 7 - 2.00 4 -1.65 4 -1.65 Middle 1.8 Middle 1.3 5 -1.69 4 -1.34 6 -1.83 5 -1.76 5 -1.71 Upper 1.8 Upper 1.4 7 - 1.43 5 -1.75 6 -1.61 6 -1.83 (available (vehicles 6 -1.83 28 November 8, 2010 PBS&J Hurricane 2000 Miller 2004 Miller 2008 Statewide Regional Ken Metcalf Reid Ewing Evacuation Analysis Model (1990 Update Evacuation Study Program Miller Model Recommendations Dec.1991 (1990 Census & 2000 Census South Florida Behavioral Analysis - Re ort 7 - 1.58 7 - 1.43 7 - 1.44 vehicles - page used in 65) evacuation - page 65) Zone 1 - 1.04 Zone 1 - 1.04 Zone 1 - 0.83 2-1.04 1-1.04 2-1.04 2-1.23 3-1.05 3-1.05 3-1.23 4-1.10 2-1.04 4-1.10 4-1.13 5-1.10 3 - 1.05 5-1.10 5-1.13 6-1.10 6-1.10 6-1.55 Number of Vehicles per 7 -1.10 4 - 1.10 7 - 1.10 7 - 1.55 Tourist Unit 5 - 1.10 6 - 1.10 7 - 1.10 % Participation of M.H. 95% 95% 95% 100% Units Zone 1 - 60% Zone 1 - 60% Would leave if Would Zone 1 - 70-90% 2-60% 2-60% mandatory leave if 2 - 70-90% 3 - 80% 3 - 80% evacuation notice mandatory 3 - 75-95% 4 - 85% 4 - 85% is given for a Cat evacuation 4 - 75-95% 5 - 85% 5 - 85% 3 Hurricane (page notice is 5 - 75-95% 6 - 85% 6 - 85% 36) given for a 6 - 75-95% 60% lower keys (1 7 - 85% 7 - 85% Key West 77% Cat 5 7 - 75-95% % Participation of &2) Lower 69% Hurricane Category 5 Storm Other Units 80% middle keys (3) Middle 74% (page 36) 85% upper keys (4-7) Upper71% Key West 89% Lower 91% Middle 90% Upper 84% Zone 1 - 86% Zone 1 - Zone 1 - 83.5% Zone 1 - 67% 2 - 71 % 84.10% 2 - 69.8% 2 - 54% 3 - 69% 2 - 66.85% 3 - 56.6% 3 - 47% 4 - 57% 3 - 58.95% 4 - 47.9% 4 - 35% 5 - 66% 4 - 45.43% 5 - 60.2% 5 - 46% % Occupancy of 6 - 65% 5 - 57.99% 6 - 67.6% 6 - 52% Dwelling Units 7 - 42% 6 - 66.37% 7 - 33.3% 7 - 27% 7 - 32.84% 2008 Estimate 83% % Participation by 95% 100% 100% 17% downward Tourists Units at Risk adjustment for evacuating by air Zone 1 - 72% 63.77% - average July 2008 Smith 2 - 64% Keys occupancy Travel Research 3 - 64% 2003-2007 Zone 1 - 82% 4 - 70% 73-78% June -July 2 - 71% % Occupancy of 45 % low occupancy 5 - 70% 45% low (peak summer 3 - 71% Tourist Units 75% high occupancy 6 - 70% occupancy months) 4 - 71% 7 - 70% 45-57% Sept - 5 - 71 % October (lowest) 6 - 77% 70.38% average 7 - 71 % 29 November 8, 2010 PBS&J Hurricane 2000 Miller 2004 Miller 2008 Statewide Regional Ken Metcalf Reid Ewing Evacuation Analysis Model (1990 Update Evacuation Study Program Miller Model Recommendations Dec.1991 (1990 Census & (2000 Census) South Florida Behavioral Analysis - Report Key West occupancy 2003- 2007 Zone 1 - 69% Zone 1 - 69% Zone 1 - 69% Key West 91 % Zone 1 - 80% 2-69% 2-69% 2-69% 2-72% 3 - 70% 3 - 70% 3 - 70% Lower 72% 3 - 79% 4-71% 4-71% 4-71% 4-80% 5-71% 5-71% 5-71% 5-80% 6-71% 6-71% 6-71% Middle79% 6-80% Vehicle Usage% 7-71% 7-71% 7-71% ° 7-80% Upper80% (% of available vehicles used in evacuation - page 65) Tourist Vehicle Usage 100% 100% 100% Zones 1 to 7 Zones 1 to 7 = % Distribution Public = 0% 0% Shelters Residents Out of County Zones 1 to 3 Zones 1 to 3 = = 5% 5% Zone 1 - 90% (Perm. Residents) Zones 4-7 = Zones 4-7 = 2 - 90% 3 - 95% 95 Friend/Relative 0% 0% 4-100% 5-100% Zones 1 to 7 Zones 1 to 7 = 6-100% = 0% 0% 7-100% Hotel/Motel Zones 1 to 3 Zones 1 to 3 = = 95% 95% Zones 4-7 = Zones 4-7 = Out of County 100% 100% 30 Florida f Department 1 Transportation !0! NW 111 th Avenue GOVERNOR Miami, Florida 33172 June 18, 2010 Craig Diamond Florida Department of Community Affairs Division of Community Planning 2555 Shumard Oak Boulevard Tallahassee, Florida 32399-2100 Re: Traffic Flow Rates for Emergency Evacuation in the Florida Keys Dear Mr. Diamond: Pursuant to your request, the Florida Department of Transportation (FDOT) was asked to provide Traffic Flow Rates for Emergency Evacuation in the Florida Keys for the Florida Department of Community Affairs' (DCA) update of the 2001 Florida Keys evacuation model. Based on our analysis, FDOT has identified "Maximum Sustainable Traffic Flow Rates per Functional Evacuation Lane" for hurricane evacuation purposes for use when conducting evacuation traffic analyses on US-1 in the Florida Keys. Please see the attached tables recommended for use in evacuation planning analyses in the Florida Keys. Table 2A identifies the existing lane configuration of US-1. Table 2B provides the maximum flow rates that could be reasonably sustained under extended periods of time for evacuation -level of demand per each segment identified. The efforts undertaken to determine these rates included a site -specific capacity study in which traffic flow data were collected and analyzed under a variety of demand conditions. A comprehensive review of traffic conditions that have occurred during other hurricane evacuations in Florida, specifically the Florida Keys, as well as in the State of Louisiana, was also conducted. Our studies incorporated data over a 10-year period since the original 2001 Keys Evacuation Study was conducted. Most importantly, this data includes observational studies of actual hurricane evacuations that have added to our understanding of traffic operations under mass evacuation demand conditions. Should you have any questions or require additional information, please do not hesitate to contact me, or Ms. Barbara Culhane, AICP, Senior Project Manager, at (305) 470-5200. Sincerely, o Aile $nB4ouV, CP District Six PLEMO Administrator www.dot.state.fl.us 0 RECYMED PAPER Craig Diamond June 18, 2010 Page 2 CC: Sandy Meyer, DEM Rebecca Jetton, DCA Jeff Alexander, NERFC Richard Ogburn, SFRPC Roman Gastesi, Monroe County Christine Hurley, Monroe County John Taylor, FDOT Vidya Mysore, FDOT Ed Ward, FDOT Gus Pago, FDOT Debora Rivera, FDOT Gary Donn, FDOT Omar Meitin, FDOT Barbara Culhane, FDOT Phil Steinmiller, FDOT 1R �� Area Miiemarkers From To Location/Description Year 2010 Configuration Lower Keys i 2.0 i 4.0 IKey West to Stock Island 4L Lower Keys ; 4.0 9.0 ;Stock Island to Big Coppitt Key 4LD Lower Keys ; 9.0 17.0 Big Coppitt Key to Sugarloaf Key 2L Lower Keys ; 17.0 22.0 ;Sugarloaf Key to Cudjos Key i 2L Lower Keys 22.0 ; 1 24.0 u ioe Key umme an ey ova 'Airport 2L Lower Keys ' 24.0 _ ' 25.0 isurnmerfand Key GOVe Airport to iSummerland Key ; 3L Lower Keys ; 25.0 30.0 'Summerland Key to Big Pine Key ' 2L Lower Keys ; 30.0 34.0 jBig Pine Key to West Summerland Keys 2L Lower Keys 34.0 I 35.2 i st summeriandeys o parns ;Harbor Keys ; 2L Lower Keys ; 35.2 36.5 15panisn RarDor eys o a is on a 1Bridge 4LD Lower Keys ; 36.5 ; 37.5 ;Bahia Honda Bridge to Bahia Honda Key 2L Middle Keys 1 37.5 i 47.0 iBahla Honda Key to Hog Key i 2L Middle Keys 47.0 48.0 '-Hog Key to Boot Key 2L Middle Keys i 48.0 i 50.2 Boot Key to Marathon 4L Middle Keys ; 50.2 50.8 jMarathon to Marathon Shores I 5L Middle Keys ; 50.8 54.0 Marathon Shores to Key Colonial Beach 4LD Middle Keys 54.0 54.5 ;Key Colonial Beach to Deer Key 4LD Middle Keys ; 54.5 58.0 {Deer Key to Grassy Key 2L Upper Keys ; 58.0 74.0 ;Grassy Key to Matecumbe Harbor 2L Upper Keys ; 74.0 80.0 ;Matecumbe Harbor to Teatable Key i 2L Upper Keys i 80.0 i 83.5 ITeatable Key to Islamorada i 3L Upper Keys ' 83.5 ' 85.6 'Islamorada to Windley Key i 2L Upper Keys ; 85.6 90.0 ;Windley Key to Plantation Key 2L Upper Keys ; 90.0 100.0 ;Tavernier Key to Newport Key 4LD Upper Keys ; 100.0 105.0 ;Newport Key to Sexton Cove 4LD Upper Keys ; 105.0 is 106.3 ;Sexton Cove to Rattlesnake Key 4LD Upper Keys 106.3 126.5 Rattlesnake Key to Card Sound Rd 2L/4L South Dade ' 126.5 , HEFT 'Card Sound Rd to HEFT i 4LD ; Upper Keys ; 106.3 ; Int CR 905 / ; ; CR 905 A (Lake Surprise to Crocodile Lake ' 2L Upper Keys ;Ocean Reef Int CR 905 ! U PP Y CR 905 A ,Tanglefish Key to Crocodile Lake 2L Int CR 905 /1 Upper Keys ; CR 905 A i US 1 i Crocodile Lake to South Miami -Dade i 2L LEGEND 2L Two-lane facility 2L/4L Two lanes with short four -lane sections for passing purposes 3L Three -lane facility (center lane is a two-way left -turn lane) 4L Four -lane undivided facility 4LD Four -lane divided facility 5L Five -lane facility (center lane is a two-way left -turn lane) Area Milemarkers From To Location/Description Suggested Maximum Sustainable Flow Rate per Hour per Functional Evacuation Lane ,er Keys 2.0 ! 4.0 ':Key West to Stock Island 900 er Keys = 4.0 9.0 IStock Island to Big Coppitt Key ! 900 er Keys 9.0 ! 17.0 ( Big Coppitt Key to Sugarloaf Key 1,100 er Keys ! 17.0 ! 22.0 ,Sugarloaf Key to Cudjoe Key 1,100 er Keys ! 22.0 ( 24.0 !Cudjoe Key to Summedand Key Cove Airport 1,100 er Keys ! 24.0 25.0 1,Summerland Key Cove Airport to Summedand Key 1,100 er Keys 25.0 ! 30.0 !Summedand Key to Big Pine Key 1,100 er Keys ! 30.0 34.0 ! Big Pine Key to West Summedand Keys f 1,050 er Keys ! 34.0 ! 35.2 West Summerand Keys to Spanish Harbor Keys 1,100 er Keys 1 35.2 ! 36.5 !Spanish Harbor Keys to Bahia Honda Bridge 1,100 er Keys ! 36.5 ! 37.5 !Bahia Honda Bridge to Bahia Honda Key 1,100 ile Keys ! 37.5 ! 47.0 Bahia Honda Key to Hog Key 1,200 ile Keys ! 47.0 ! 48.0 !Hog Key to Boot Key ! 1,100 ile Keys ! 48.0 ! 50.2 ! Boot Key to Marathon ` 900 ile Keys ! 50.2 ; 50.8 (Marathon to Marathon Shores ! 900 ile Keys ! 50.8 ( 54.0 !Marathon Shores to Key Colonial Beach ! 900 ile Keys 54.0 ! 64.5 !Key Colonial Beach to Deer Key I 900 Ile Keys I 54.5 I 58.0 (Deer Key to Grassy Key I 1,100 er Keys ! 58.0 74.0 !Grassy Key to Matecumbe Harbor ! 1,100 er Keys 74,0 1 80.0 1 Matecumbe Harbor to Teatable Key 1,100 er Keys 80.0 I 83.5 jTeatable Key to Islamorada 1,100 er Keys ! 83.5 ! 85.6 ( Islamorada to Windley Key ! 1,100 or Keys ! 85.6 1 90.0 IWindley Key to Plantation Key t 1,100 or Keys I 90.0 I 100.0 'Tavernier Key to Newport Key ! 900 er Keys ! 100.0 ; 105.0 'Newport Key to Sexton Cove 900 er Keys ! 105.0 ! 106.3 !Sexton Cove to Rattlesnake Key ! 900 er Keys 106.3 ; 126.5 !Rattlesnake Key to Card Sound Rd 1,200 Ih Dade ! 126.6 ! HEFT Card Sound Rd to HEFT ! 900 er Keys ! 106.3 Int CR 905 ! CR 905 A !Lake Surprise to Crocodile Lake ! 1,100 I i Int CR 905 er Keys Ocean Reef ! CR 906 A !Tanglefish Key to Crocodile Lake 1,100 ! Int CR 905 us 1 Lake to South Miami -Dade NOTES A Functional Evacuation Lane has a pavement width of at least 10 feet The above flow rates are maximum values that are expected to be sustained for extended periods (more than 8 hours). During night conditions, these flow rates may be lower than the ones shown above. 100 Florida Department of Transportation CHARLIE CRIST 1000 NW 1 1 1 th Avenue STEPHANIE C. KOPELOUSOS GOVERNOR Miami, Florida 33172-5800 SECRETARY October 28, 2010 Christine Hurley, AICP Monroe County Growth Management Division Director 2798 Overseas Highway, Suite 400 Marathon FL 33050 Dear Christine: This is in response to your e-mail from October 27, 2010 regarding the Florida Keys Hurricane Evacuation Update. As requested in your e-mail, attached are the following items: 1. Attachment 1 contains Technical Memoranda prepared by Dr. Brian Wolshon, PE, of Louisiana State University and Joaquin Vargas, PE, of Traf Tech Engineering, Inc., for the Florida Department of Transportation (FDOT) District Six entitled The Florida Keys Site -Specific Capacity Study Technical Memorandum (dated April 21, 2010) and the Maximum Sustainable Evacuation Traffic Flow Rates for Hurricane Evacuation Analysis Purposes Technical Memorandum (dated June 17, 2010). 2. Attachment 2 contains a spreadsheet with the results of hourly traffic counts conducted from January 28-31, 2010 at three locations. These traffic counts were recorded for the Florida Keys Site -Specific Capacity Study dated April 21, 2010. As indicated in the attached count sheet, the maximum northbound hourly volume recorded on US-1/SR 5 near Mile Marker 106 (US 1/SR 5, just south of the 18-mile Stretch/CR 905 intersection) was 1,332 vehicles per hour (two lanes), or 666 vehicles per hour per lane. The subject count occurred on Sunday, January 31, 2010 between 2:00 PM and 3:00 PM. 3. Information related to signal operation during hurricane evacuation: Other professionals have previously suggested that all Monroe County traffic signals should operate in the "flashing" mode during hurricane evacuation conditions. This is not advisable at critical intersections located along US 1/SR 5 within the Florida Keys. As documented in the Florida Keys Site -Specific Capacity Study dated April 21, 2010, within the Key Largo area, it is preferred that the traffic signals located at Atlantic Boulevard and at Tarpon Basin Drive remain operational during hurricane evacuation conditions for the following reasons: o Nearly 20 percent of the evacuating vehicles will enter US 1/SR 5 from Key Largo. As such, drivers from this area will need adequate gaps to permit safe merging into the outbound US 1/SR 5 traffic stream The -US 1ISR _5 segment between We Markers_99.5_ and _106.3 will _carry --the heaviest traffic volumes of the entire Florida Keys evacuation network www.dot.state.fl.us www.dot.stateftus Christine Hurley October 28, 2010 Page 2 o By maintaining full operation of the traffic signals located near Mile Markers 100 and 101, gaps will be created along US 1/SR 5 which will benefit all evacuating vehicles entering the main highway from the numerous side streets. (Similar reasons apply at other locations throughout the Florida Keys). To minimize the effects of the traffic signals on the evacuating traffic flow along Overseas Highway, the two simulated traffic signals were optimized so that most of the green time was allocated to US 1/SR 5. That is, the assumed signal operating plan permits one vehicle tuming left from US 1/SR 5 and up to two vehicles entering Overseas Highway from the side street, per signal cycle. This is an assumption that significantly benefits the evacuating flow along the US 1/SR 5. 4. Attachment 3 contains tables with recommended number of Functional Evacuation Lanes for the following scenarios: a. Completed projects from Table 18 of the 2001 Keys Evacuation Study (2001 Study) b. Projects under construction from Table 18 of the 2001 Study c. Projects funded in current work program from Table 18 of the 2001 Study 5. The table depicting FDOT Work Program Roadway Projects for the Florida Keys Evacuation Improvements as requested are hurricane related evacuation improvement only, as recommended in the 2001 Florida Keys Hurricane Evacuation Study. If you have any questions please contact me at (305) 470-5200 or by e-mail at aileen. boucleCa-dot. state.fl. us. Sikee y Aucle, AICP District Planning and Environmental Administrator CC: Rebecca Jetton, DCA Barbara Culhane, FDOT Reid Ewing, PhD Ken Metcalf Enclosures 1 d WE', 3 lid 3 � s s OR] ] s B ',AiN rl; i t t, ; € � 0 WV E t Est 6 9 A aE,3wig 9 9 9 t Nq b �x st-� F - CIA� x an tuS � c ti Sti y t _ � it ° i R LL { � S M " c Mr sY itsY s c gas, �xxs gWAN s iss \' � g WA gas�ssssssE s � x e � 1 ffix ll 5`�. ox���e e;se 5� - _�_lax. 't£tt, ��A mugsx