Wess Gar Cag Kes. aoe, Uh Kf. CER C 2 SC a o wD) P YG. Gy, TECHNICAL REPORT CERC-86-8 US Army Corps FREQUENCY OF COASTAL FLOODING AT Sima Ss ROUGHANS POINT, BROAD SOUND, LYNN HARBOR, AND THE SAUGUS-PINES RIVER SYSTEM by Thomas A. Hardy, Peter L. Crawford Coastal Engineering Research Center DEPARTMENT OF THE ARMY Waterways Experiment Station, Corps of Engineers PO Box 631, Vicksburg, Mississippi 39180-0631 September 1986 Final Report Approved For Public Release; Distribution Unlimited DOCUMENT LIBRARY Woods Hole Oceanographic Institution Prepared for US Army Engineer Division, New England Waltham, Massachusetts 02254-9149 Destroy this report when no longer needed. Do not return it to the originator. The findings in this report are not to be construed as an official Department of the Army position unless so designated by other authorized documents. The contents of this report are not to be used for advertising, publication, or promotional purposes. Citation of trade names does not constitute an official endorsement or approval of the use of such commercial products. Unclassified SECURITY CLASSIFICATION OF THIS PAGE Form Approved REPORT DOCUMENTATION PAGE OMB No. 0704-0188 Exp. Date: Jun 30, 1986 Ja. REPORT SECURITY CLASSIFICATION 1b. RESTRICTIVE MARKINGS Unclassified 2a. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION / AVAILABILITY OF REPORT Approved for public release; 2b. DECLASSIFICATION/ DOWNGRADING SCHEDULE distribution unlimited 4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT NUMBER(S) Technical Report CERC-86-8 6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL | 7a. NAME OF MONITORING ORGANIZATION USAEWES, Coastal (if applicable) Engineering Research Center WESCV 6c. ADDRESS (City, State, and ZIP Code) 7b. ADDRESS (City, State, and ZIP Code) PO Box 631 Vicksburg, MS 39180-0631 8a. NAME OF FUNDING / SPONSORING 8b. OFFICE SYMBOL || 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER ORGANIZATION US Army (If applicable) Engineer Division, New England 8c. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS PROGRAM PROJECT TASK WORK UNIT 424 Trapelo Road ELEMENT NO. | NO NO ACCESSION NO Waltham, MA 02254-9149 11. TITLE (Include Security Classification) Frequency of Coastal Flooding at Roughans Point, Broad Sound, Lynn Harbor, and the Saugus-Pines River System 12. PERSONAL AUTHOR(S) Hardy, Thomas A., Crawford, Peter L. 13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year, Month, Day) | 15. PAGE COUNT Final report rromMay 1984 tec 1985 September 1986 98 16. SUPPLEMENTARY NOTATION Available from National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161 Ze COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number) GROUE SUEKEIROU? Coastal flooding Wave overtopping Storm surge Flood frequency Stage-frequency 19. ABSTRACT (Continue on reverse if necessary and identify by block number) This report describes the establishment of frequency curves for water levels caused by the combination of tide, storm surge, and waves for a coastal area just north of Boston. The project procedure involves the conjunctive use of five modeling components, including numerical storm surge, numerical wave propagation, physical wave overtopping, flood routing, and probability models. At Roughans Point where flooding is caused by the overtopping of seawalls by storm waves, all five models were necessary. Multiple combinations of possible seawall-revetment structures were modeled. Major differences among the combinations were evident at the lower return periods with the combinations of a wide berm revetment and a cap on the existing sea- wall for the east wall of Roughans Point providing the greatest protection. At higher return periods the protection differential offered by the various structure combinations tended to diminish. For still-water levels and wave conditions of a Standard Project (Continued) 20. DISTRIBUTION / AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION (1 UNCLASSIFIED/UNLIMITED © same As RPT Optic users Unclassified 22a. NAME OF RESPONSIBLE INDIVIDUAL 22b. TELEPHONE (Include Area Code) | 22c. OFFICE SYMBOL DD FORM 1473, 84 MAR 83 APR edition may be used until exhausted SECURITY CLASSIFICATION OF THIS PAGE All other editions are obsolete Unclassified AUN UU L 0 0301 0091254 4 20. ABSTRACT (Continued) Northeaster all structure combinations tested would be ineffective at pro- tecting the interior of Roughans Point. Tests were conducted to determine a structure height for the north wall. These tests indicated that significant overtopping did not begin until the north wall structure was lowered below 13 ft National Geodetic Vertical Datum (NGVD). Since the existing height of the north wall is above this level at several sections, it is recommended that the revetment height be set at 13 ft NGVD with the wall height set so that there is a transition between the existing wall heights. For areas where stage-frequency curves are presented for the still-water level resulting from the combination of storm surge and astronomical tide, only the storm surge and probability models were necessary. These areas in- clude both open coast and estuarine locations. For areas flooded by the still-water level, results of the modeling indicated that the whole study area floods to approximately the same level. Flood levels are efficiently conveyed through the inlet and throughout the flood plain of the Saugus-Pines River system. Inside the inlet, there is a small gradient in the still-water level, rising from north to south, which results from local setup caused by north to northeast wind directions which predominate during storm conditions. This local wind setup results in flood levels inside the inlet which differ by one- half to three-fourths of a foot during the more severe storm events. Outside the river system in Broad Sound a smaller north-south gradient exists with differences of only a few tenths of a foot resulting. Data collected by the US Army Engineer Division, New England, after completion of the modeling indicated that losses do occur as flood levels propagate upstream of the Fox Hill Drawbridge on the Saugus River and upstream of the Highway embankment on the Pines River. Stage-frequency curves for these areas were adjusted to accommodate these additional data. The curves were lowered 0.3 and 0.5 ft at the lower return periods for upstream Saugus River and Pines River locations, respectively. Reductions were reduced for higher return periods because higher flood levels would provide greater access of floodwaters to these areas. The setup and operation of all models, except the physical model, are de- scribed. The method of constructing stage-frequency curves is explained, and estimates of the error involved in each of the processes are discussed. The final products are curves which relate flood stage to frequency of occurrence for several possible structures at Roughans Point as well as for several coastal and river areas. PREFACE The US Army Engineer Division, New England (NED) requested the US Army Engineer Waterways Experiment Station (WES) Coastal Engineering Research Center (CERC) to conduct numerical and physical model studies to determine the frequency of flood levels at Roughans Point and at other coastal areas in Revere, Saugus, and Lynn, Massachusetts. The studies were conducted prin- cipally to provide greater confidence in the flood protection plan for Roughans Point as presented in the planning report (NED 1983) and were part of a larger study, "Continuing Planning and Engineering Studies for Roughans Point," provided for under the 12 September 1969 Southeastern New England authorization of the US Senate Committee on Public Works. A small funding contribution came from Revere Backshore planning studies conducted under the Same authority. This report contains the results of the numerical investigations con- ducted between May 1984 and December 1985. Close consultation and coopera- tion were maintained between CERC and NED throughout the study, and the efforts of Mr. Charles Wener, NED, were particularly important in its suc- cessful completion. Work was performed by personnel of the Research Division (CR), CERC, under the direction of Dr. James R. Houston, former Chief, CR. Mr. Thomas A. Hardy, Coastal Processes Branch (CR-P), was the Principal Investigator for this study under the direction of Mr. H. Lee Butler, former Chief, CR-P, and current Chief, CR. Mr. Hardy was responsible for the probability modeling, storm surge modeling, flood routing, and synthesis of the total modeling ef- fort. Mr. Peter L. Crawford, Coastal Oceanography Branch (CR-0), was respon- Sible for the wave modeling. Mr. Crawford worked under the direction of Dr. E. F. Thompson, Chief, CR-O. Upon completion of the study, Chief and Assistant Chief, CERC, were Dr. Houston and Mr. Charles C. Calhoun, Jr., respectively. This report was edited by Ms. Shirley A. J. Hanshaw, In- formation Products Division, Information Technology Laboratory, WES. COL Allen F. Grum, USA, was the previous Director of WES. COL Dwayne G. Lee, CE, is the present Commander and Director. Dr. Robert W. Whalin is Technical Director. CONTENTS PHONO 5 660600000000000000000000000000000000005000000000800006 CONVERSION FACTORS, NON-SI TO SI (METRIC) UNITS OF MEASUREMENT PART I: NPA ODING MMOs oo ocooodo Dodd DDD DODD DOOD DOU OSODDOOOONN NEAMAMOMCEAV ao ooo oO D DODO OD OOOO DD DOOD DUO OOO ODDO OOD DOODNNDS Overview of Project Technique.................2-e-eeeee- Ojeeialeerlon. Oe INNO. 50560050500000000000000000000000000 PART II: PROVSVUSIUEIINE WOOD. oo oo000c0gd00 DDD DOO DDDD ODD OGD0000 Choosing Storm Surge Time-Histories..........ceesseeeeee Creating Synthetic Surge Plus Tide Events............... Selecting: Events) Co Mod eis « «oie ejeienerelone olckoleelaieyclensiel elo PART LIT: STORM SURGE) PEUS TIDE SIMULATION 2 occ ee (Cyortal IWIN VININGS 6 co00cGb00C DODO ODDO DOD DU OOD OORODDDOD0000 iiljoyol JROPONMR 6.6.56.0.010.0000000000000000005000000000000000000 DEWEY COMMOTION. 6 66.0060000600000000000000000000000000000 Model Cadhilbrca\talomlersepcmenorneoleielcieichnedonel olcheNcionelh-Nchloncncn Monon ieNel= Wiloyoleul WeraliPikeeNeiOlals 5 5 o.0665000000000060000000000000000000 Simulation of Event Ensemble by the Hydrodynamic Model.. PART eh ciWAVEmMODEISUN Gipepepieteieienelseneeerelenelensieicrencr Rone KoN-Nonoerelenevensions AUS) lnleteavoyols} evavol IDENCENS 6o606000000000000000000000000000000 Phasemishsia Method olllopuypierrericnernerartcrnencn nel eneioncnenclennenenenewerenenoncne Use of Phase III Methodology for Broad Sound Wave Cilia tee: SiMulensieyass oo66600000000000000000000000000000 Wave Climate Simulations for Broad Sound................ ESCUBEDBRESWUUStenarcietel uetstchel erelenorct roiled) clone leen nel cticnenerer none orale Locally Generated Waves in the Lee of Nahant Peninsula.. PART V: FLOOD STAGES FOR THE INTERIOR OF ROUGHANS POINT.... Overtopping Rate Calculation..............ccccescccesees JaAl@rexol MNereMNo Goo cocdogcddo dogg OOO DU GOO UOO0daD0G00000000 eeceee ee ee o eoceeeoe eee o eeceee ce eo eecee ee eee o eocee ee eee o eoceeee eee oe eceoeceeee eee Simulation of the Event Ensemble by the Flood Routing Model....... NORche Wally MeStS2 045 Bua) cleyevereeneuedeneencne ioroxeveuepereucuevovsrcuenersucweloreuere PART VI: STAGE=BREQUENGYSCURVESMermmicieiomiciociie nici PART Avie RESULTS SANDMCONCEUSIONSirersererelcucicnercionen cacti neicneicreietcrorene OPRAREINS JONG 5 odGb0000 cdo ODDO DdODDD DDD ODDDDDDDNDDDODNNO Sisal Wat ers locationSireaneccnercieirelcicickersnenorncirenenononeineneroncne Estimating Error in the Frequency Curves................- Determining Error Bands for the Selection Process....... Assessing the Impact of the Standard Project Northeaster (CloyayeNNS WINS} 56.6.6. 0'6.06.000000000540000000000.00000000000000 REFERENCE Sipenavatstechetcivesonenerelel clot oelen homcnot verano eneiolor Meno temeneltomon Meher omen en etch is eececeeeeeee eco e eee e ee CONVERSION FACTORS, NON-SI TO SI (METRIC) UNITS OF MEASUREMENT Non-SI units of measurement used in this report can be converted to SI (metric) units as follows: Multiply By To Obtain acres 4 046.873 square meters cubie feet per second 0.02831685 cubic meters per second feet 0.3048 meters knots (international) 1.8532 kilometers per hour miles (US statute) 1.609344 kilometers square miles (US statute) 2.589988 square kilometers FREQUENCY OF COASTAL FLOODING AT ROUGHANS POINT, BROAD SOUND, LYNN HARBOR, AND THE SAUGUS-PINES RIVER SYSTEM ee PART I: INTRODUCTION 1. The study area was located in the cities of Revere, Lynn, and Saugus, Massachusetts, which are immediately north of Boston. Roughans Point (Figure 1) is a 55-acre* residential area which is below the elevation of a spring tide at many locations. Seawalls along both the northern and eastern boundaries offer some protection against coastal flooding. However, damage resulting from flooding caused when waves overtop the seawalls is a frequent occurrence. The Saugus and Pines Rivers join just before passing under the General Edwards Bridge and out into Broad Sound. The lower 2,500 acres of the drainage area just behind Revere Beach are mostly river channel and marsh. This area borders residential, commercial, and industrial areas, many of which are at an elevation only a few feet greater than the elevation of the maximum astronomical tide. Flooding is caused by the inundation of low lying areas by the combination of astronomical tide and storm surge. The Revere Beach-Point of Pines-Lynn Harbor region is made up of recreational beaches, residential and industrial land protected by seawalls, and harbor areas. Flooding results from overtopping of seawalls and dunes by storm waves. Figure 2 is a map of the study area vicinity showing the above locations. 2. The desired products of this project are stage-frequency curves which relate the elevation of floodwaters to the average waiting time between floods of equal or greater severity. The ordinate of these curves is stage, measured in feet above the National Geodetic Vertical Datum (NGVD), and the abscissa is return period expressed in years. The primary goal of this study initially was to provide flood frequencies at Roughans Point where flooding is caused by the overtopping of seawalls by storm waves. The numerical model efforts needed to predict waves and water levels at Roughans Point could also predict these quantities at nearby locations. Therefore, the scope of the project was expanded to provide flood frequencies for the Saugus-Pines River * A table of factors for converting non-SI units of measurement to SI (metric) units is presented on page 3. Figure 1. Location of study area LYNN AUG LYNN rs ‘ny Hang Weg SAUGUS a LOGAN “:) AIRPORT Figure 2. Study area vicinity System, as well as wave and water level information, and techniques which could be used for future overtopping studies at Point of Pines and Lynn Harbor. The study was then divided into two main sections, determined by whether the cause of the water levels was due to wave overtopping or combined surge and tide. Roughans Point was the only location where stage-frequency curves were generated for flooding resulting from wave overtopping. For the Saugus-Pines River System, flooding results from the inundation of low lying areas by the combination of storm surge and astronomical tide. Even though flooding at Revere Beach, Point of Pines, and Lynn Harbor is caused mostly by Wave overtopping, only the still-water level frequency will be reported be- cause the present study did not include investigation of overtopping for these areas. Wave overtopping for these areas will be estimated by US Army Engineer Division, New England (NED), in other studies using techniques and data devel- oped by US Army Engineer Waterways Experiment Station (WES) Coastal Engi- neering Research Center (CERC) for the present study. Areas where the stage- frequency curves are based upon combined surge and tide levels, but include no wave effects, will be referred to as still-water level locations. Terminology 3. To avoid excessive repetition and to provide greater clarity, the following terms are defined for use throughout this report. Event Storm plus tide MSL Mean sea level NGVD National Geodetic Vertical Datum (formerly called mean sea level datum of 1929) Northeaster Extra-tropical storm Stage Elevation of the still-water level above NGVD Still-water Level Elevation of surge plus tide water surface Storm The historical meteorology (wind, waves, and surge) independent of the tide with which it actually occurred Surge Storm-induced component of still-water level Tide Astronomical tide Overview of Project Technique 4, The establishment of frequency curves required the conjunctive use of several modeling components. At Roughans Point the combined use of prob- ability, numerical storm surge, numerical wave, physical, and flood routing models was required to produce the stage-frequency curves. Whereas, for the still-water locations (Saugus-Pines River and Revere Beach-Lynn Harbor areas) only the probability and numerical storm surge models were required. The following is a brief description of each model. 5. The probability model was designed to complete four tasks: select events for simulation by the other models, assign probabilities to these events, create stage-frequency curves, and determine a measure of confidence in the final results. The numerical storm surge model simulated the storm plus tide events producing a time-history of still-water levels at specific locations throughout the study area. A numerical, spectral wave model simu- lated the wave field which accompanied each of the events simulated by the storm surge model. Also, a monochromatic wave model estimated the locally generated waves which were not considered in the spectral model. The wave parameters of height, period, and direction were calculated at selected sites throughout the study area. The physical model determined coefficients for an overtopping rate equation by testing multiple combinations of water level and spectral wave characteristics for several existing and proposed structures at Roughans Point. The physical modeling is not fully described in this report. (For complete details of the physical modeling see Ahrens and Heimbaugh (in preparation). The flood routing model calculated the maximum stage in the interior of Roughans Point caused by each event. Maximum stage was determined after outflows from drainage, pumping, seepage, and weir flow over low lying boundaries were considered. 6. Figure 3 is a flow chart which depicts the conjunctive use of the above models for the establishment of stage-frequency curves. Basically, the probability model selected and assigned probability to the surge-tide-wave events simulated. Then, the surge model simulated the still-water level. At this point stage-frequency curves were generated for the still-water loca- tions. To develop the flood levels caused by wave overtopping at Roughans Point, the wave, physical, and flood routing models were necessary. The wave model simulated the parameters, height, period, and direction. The output of enbtuyoe, yoefoud jo 4queyo MoTA “€ aun3Ty NOILVNDA NOILVINI1V9 SAIYOLSIH-AWIL JLVY INIddOLYAAO INIddOLYAA0 SYALIWVEVd JAVM T3001 WAISAHd 1300 JNILNOY GO0T4 SNOILVIO1 YALVM-TIHLS INIOd SNVHONOY $14A41 G0014 YOUU WNWIXVA J1aVd0ud NOILVYANAS AAYNO AONANDAYS SIAUNI AININOIYS-IOVLS SAIYOLSIH-AWIL TdAd1 YALVM INAWNOSISSV ALNMIGV80"d NOILOA1SS IN3A4 1I00N JOUNS WHYOLS 1300 ALNIGVa0ud the two numerical models (surge and wave) were the main inputs to the physical model's overtopping rate equation which produced the overtopping rate for each required time-step. The water volume due to overtopping was then routed through the Roughans Point area, and a maximum stage was calculated for each event. Finally, a stage-frequency curve was created for flood levels induced by wave overtopping. Organization of Report 7. This report is structured as follows. Part II is a description of the probability model. Modeling of the surge plus tide events is discussed in Part III, including calibration and verification of the storm surge model. Part IV is a description of the numerical wave modeling. The methods for cal- culating the overtopping rate time-histories and routing the flood through Roughans Point are discussed in Part V. The construction of stage-frequency curves is explained in Part VI. Part VII contains discussion of the results, including an estimate of the error in the stage-frequency curves. Because of the large volume of results generated by the numerical models, time-histories of ocean water levels, waves, winds, overtopping rates, and Roughans Point in- terior flood levels are not provided in this report but were given to NED on computer tape. 10 PART II: PROBABILITY MODEL 8. Unlike the physical model which simulates a physical process with physical operations and the numerical models which simulate physical process- es with mathematical operations, the probability model does not simulate a physically realizable entity. The title 'model' is used for symmetry with the other components of this project. The probability model is essentially an assemblage with four specific tasks: select events for simulation by the other three models, assign probabilities to these events, create stage- frequency curves, and determine a measure of confidence in these curves. 9. Ideally, there would be a long historical data record of the desired quantity at the desired location (for example, 100 years of overtop- ping data at Roughans Point). For this ideal case modeling would not be necessary. An overtopping rate frequency curve could be created using well- established statistical techniques which can be found in any hydrology text. However, as is usually the case, sufficient data records for the quantities of interest were not available. Therefore, three separate modeling efforts, a physical overtopping model, a numerical storm surge model, and a numerical Wave model were implemented to overcome the lack of data. 10. There are several possible approaches in establishing frequency curves where the scarcity of data in the immediate study area requires a mod- eling approach. The two most common are called the historical method and the joint probability method (JPM). In the historical method, a series of histor- ical events is recreated with the pertinent data being saved in the necessary locations. In effect, it is like operating a time machine with the hindsight to know what data to collect and where to collect it. Probability is assigned to each event by a standard ranking method. For the JPM, the storm type is parameterized. For example, hurricane wind fields can be defined by three Parameters, central pressure deficit, radius to maximum winds, and forward speed. Then, an ensemble of synthetic events is simulated representing those events which are possible in the study area. Probability is assigned to in- dividual events by assigning probabilities to parameter values which determine that event. If the parameters are independent, then the probability of the event would be the product of the probabilities of the component parameters. Several studies have been conducted using the above two methods, including Meyers (1970) and Prater, Hardy, and Butler (in preparation). For the present 11 study, since hurricanes do not significantly contribute to stage-frequencies in the project area, and since northeasters are difficult to parameterize, a modification of the historical approach was used. Historical storm surge time-histories were combined randomly with tide time-histories to produce syn- thetic event water level time-histories. Probabilities were assigned using data from a nearby tide gage. This process is explained in detail in the following paragraphs. Choosing Storm Surge Time-Histories 11. Regardless of the approach selected, data in the vicinity of the study area are essential for identifying inputs to the numerical modeling and for assigning probabilities. This project was fortunate in having convenient sources for the necessary data. The National Ocean Service's (NOS's) Boston tide gage has been in continuous service since 1922. This gage is located at Commonwealth Pier in Boston Harbor which is less than 5 miles from the study area. Wave hindcast information was available for deep water adjacent to the study area from the WES Wave Information Study (WIS). Hourly wind data were available from Logan International Airport which is less than 5 miles from the study area. The 20-year period from 1956-1975 was chosen from which to gather data for use in the numerical modeling. This period was selected because information was available from the above mentioned sources in all the necessary data categories: water level, wind, and wave. 12. By defining storm surge as the difference between measured water level and predicted tide, a partial duration series of storm surge time- histories (26 storms) was extracted from the Boston tide gage data. A minimum value of the maximum surge, 2.5 ft, was used to define those storms which had a reasonable probability of causing significant flooding. If surges much below the 2.5-ft level were combined with possible tides, it would be unlikely that any of the resulting events would be selected as one of the relatively small number of events to be modeled (only 150 events with water levels from 7.9 to 11.2-ft NGVD were selected). The value of 2.5 ft was chosen using the following guidance: with this surge level, only 5 percent (Harris 1981) of the hourly tide heights are high enough so that the combined surge plus tide would be greater than 7.9 ft NGVD. The combination of surge and tide and the selection of the events to be modeled are explained later in this report. 12 13. Two additional storms from outside the 1956-1975 period were in- cluded in the storm ensemble: 29 November 1945 and 6 February 1978. All the necessary data were obtainable for these storms which caused the first and second highest surges recorded at Boston. Furthermore, the February 1978 event caused the highest still-water level (10.3 ft NGVD) on record in Boston. Adding these two storms helped ensure the top end of the storm en- semble was representative of what could occur at Boston. Therefore, a total of 28 storms was chosen to represent the surge time-histories which are possible at Boston. Table 1 contains a list of these storms and their maxi- mum surges. The maximum surges listed in Table 1 might differ slightly from maximum surges derived elsewhere. There are two reasons for these small discrepancies. First, great care was taken to use a set of tidal prediction constituents which best fits the tidal signal at Boston. With the large tidal range at Boston, slight errors in phase could cause significant errors in the calculated surge. Five separate sets of constituents received from NOS were tested, and the set of constituents with the best fit was used for these calculations. Second, often the maximum surge in historical storms occurs at low water because of the increase in surge with decreasing water level given constant wind speed and direction. Since the surges needed to be independent of their historic tide, the surge time-histories were edited by eye to remove 12-hour oscillations caused by this shallow-water effect. Table 1 Historical Storms Chosen to Represent Possible Surges at Boston Storm Maximum Surge Storm Maximum Surge No. Date ft No. Date ft 1 11-30-45 4.8 15 4-13-61 44 2 1-9-56 303 16 3-7-62 n> 3 3-16-56 353) 17 12-6-62 Qo 4 4-8-56 2.6 18 2-19-64 2 si 5 1-8-58 2.9 19 1-23-66 3o 6 1-15-58 Zell 20 1-30-66 3.6 7 2-16-58 356 21 12-25-66 2.9 8 3-15-58 2.8 22 2-9-69 3.4 9 3-21-58 3.2 23 12-27-69 3.2 10 4-2-58 2 Tl ay 2-4-72 2.9 11 12-29-59 2.6 25 2-19-72 4.0 12 2-19-60 2.5 26 11-9-72 2.8 13 3-4-60 Bail 27 12-16-72 3.2 14 1-20-61 3.4 28 2-6-78 4.7 Creating Synthetic Surge Plus Tide Events 14. Since the tidal range at Boston (mean range--9.5 ft and maximum range--14.6 ft) is much larger than the largest recorded surge (approximately 5 ft), the tide is a very important component of the total water level. Rath- er than numerically model the relatively small sample of historical events (surge plus tide), synthetic events were created by combining the historical storm surge time-histories with possible tide time-histories. The basic as- sumption behind this technique is that the surge time-history (edited to re- move the shallow-water effect) of any storm is independent of the tide with which it occurs. In other words, the phenomena which cause tides are not related to the phenomena which cause storms. Therefore, a storm may occur with any tide that is possible during storm season. 15. Using tidal constituents from NOS analyses of the Boston tide gage, hourly tide heights for the winter season were predicted. The period from 15 October to 30 April was chosen as winter season, and 19 years of this seasonal record were generated to simulate a tidal epoch. Combining the 28 surge time-histories with every possible tide time-history during this tide series would result in more than 2.5 million combinations. Obviously, it would be economically impossible to simulate all these possibilities. Fur- thermore, it is not necessary to simulate a large percentage of the possi- bilities in order to adequately represent the population. In order to form a representative sample of the total ensemble, a random selection process was devised. 16. The 28 surge time-histories were combined with a large number of tide time-histories. Each of these synthetic surge plus tide time-histories Was created from storm and tide time-histories by randomly choosing a starting point in the tide series, matching this point to the start of a storm, and adding the tide and surge levels at each hour for the length of the storm. The resulting large number of possible event time-histories served as the data set from which events were randomly selected for simulation by the numerical models. Each of the 26 storms, in the 20-year partial duration series of surge, was combined with 500 tide time-histories chosen at random from the 19-year tide series. Each of these storms was considered to have an equal likelihood of occurrence (each storm did, in fact, occur during the 20 years). The two additional storms (1945 and 1978) were not part of the 20-year partial 14 duration series and, therefore, did not have the same likelihood of occurrence as the other 26 members in the ensemble. Therefore, these two extra storms were combined with a fewer number of tides. To determine the number of events which should be formed using these two storms, the following simplified analy- sis was used. The 1945 and 1978 storms had the first and second largest surges in a 58-year annual series (the length of available data). Assuming a Weibull plotting position formula, p = m/N+1, where m is the rank and N = 58, the 1945 and the 1978 surges would have frequencies of 1/59 and 2/59, respec- tively. Assuming the other 26 members of the storm ensemble to have fre- quencies of 1/20, and using the ratios of these frequencies, the 1945 storm was combined with 170 tides and the 1978 storm with 340 tides. For example, (1/59) / (1/20) x 500 = 170 . This analysis is not rigorous from a statis- tical standpoint and was done primarily to prevent the two storms with the strongest winds and largest waves from being overrepresented at low and medium water levels. Approximately 13,500 possible surge plus tide time-histories resulted from this process (26 x 500 + 340 + 170). Events to be simulated were selected from this file of possible surge plus tide time-histories. Selecting Events to Model 17. A flood-causing event is multidimensional. The severity of the damage caused by the event is determined by several factors, among which are the magnitude and duration of winds, waves, and water levels. Because of the difficulty of ranking multidimensional entities, as well as the lack of avail- able data for doing so, it is necessary to reduce the dimensionality. There- fore, only one dimension, maximum still-water level, was used to measure the severity of an event. This criterion was chosen for two main reasons. First, it was deemed the most important; and, second, there was a large volume of available data. NED has established a stage-frequency curve (Figure 4) at (NED 1983) relating maximum still-water level with its frequency of occurrence the Boston NOS tide gage. This stage-frequency curve was used as the basis for both event selection and the assignment of probability to simulated events. 18. Based upon previous experience (Prater, Hardy, and Butler, in prep- aration) it was estimated that by simulating 50 events the frequency of still- water level would be accurately represented throughout the study area. 15 STILLWATER ELEVATION (FT., NGV.D.) fe) Iygrel bh Si lly ike! 5 10 20 30 40 50 60 70 80 9390 95 98 99 PERCENT CHANCE OF OCCURRENCE PER YEAR Figure 4. NED stage-frequency curve for Boston still-water level would be accurately represented throughout the study area. However, the extra variables involved in simulating waves and wave overtopping volumes would cause added uncertainty in the final frequency curves at Roughans Point. Therefore, it was decided to simulate 150 events in order to increase the confidence that the frequency curves based upon overtopping calculations were accurate. The 150 events were selected and simulated, and then frequencies were calculated in three separate sets, each containing 50 events. This was done to establish a measure of confidence in the selection procedure. This confidence calculation will be explained in Part VI. 19. The selection process involved four steps. First, the stage incre- ments, for which simulations were to be performed, were chosen. As previously mentioned, the highest still-water level on record for the Boston area is 10.3 ft NGVD, which occurred during the February 1978 northeaster. As pre- dicted by the NED curve, the 500-year level is 11.2 ft NGVD, and the annual level is 7.9 ft NGVD. Events were selected to duplicate the NED stage- frequency curve below the 500-year level at the Boston gage. Therefore, given the small range in elevation and choosing three sets of 50 events, selections were made every 0.1 ft from 7.9 to 10.4 ft and every 0.2 ft from 10.4 to 11.2 ft. Next, the number of events to be selected at each stage increment 16 Was decided. The results of these first two processes are shown in Table 2. Examining Table 2, it can be seen that more events were selected for the lower range of water levels (=8 ft NGVD) than were selected for the higher water levels (above 10.5 ft NGVD). This was done for two reasons. First, the prob- ability mass representing the lower part of the NED curve will be much larger than the probability mass representing the higher portion of the curve. This is caused by the logarithmic nature of the frequency of water levels. Experience has shown that frequency curves are more easily constructed when the probability masses assigned to simulated events vary as little as possi- ble. For example, the probability mass per year associated with a 0.1-ft in- crement located at 8.0 ft on the NED curve is 0.14; whereas, the probability mass per year for a 0.2-ft increment at 11.2 ft is 0.00035. Therefore, more events were selected at the lower return periods to divide this large prob- ability mass into smaller segments. Secondly, especially when considering overtopping, events formed from many more combinations are possible at the lower stages (large surge plus low tide plus medium waves, small surge plus medium tide plus large waves, etc). At the higher stages fewer combinations are possible (large surge plus large tide plus large waves). Consequently, the higher end of the curve can be represented by fewer events than can the lower end. 20. Choosing the stage increment sizes and the number of events selected for simulation from each increment is a subjective decision. This decision is based on the range of stages to be represented, the largest differences in probability tolerable for accurate curve generation, and the financial constraints on the number of events that can be simulated. Unfor- tunately, the only sure way to determine if the decisions are correct is to view the results. Therefore selections are made, and the goodness of these decisions is reflected in the error bands presented in Part VIII. 21. The third part of the selection process is the actual selection of events. The 13,500 possible events, created by combining the storm surge with tide, were ranked by the maximum water level that occurred during the surge plus tide time-history. At each of the stage increments shown in Table 2, events were randomly selected from the portion of the 13,500 events with max- imum water level equal to that height increment. This was done independently for each of the three sets of 50 events. Although these maximum water levels are for the Boston NOS tide gage, events selected for simulation in the study 17 Center of Center of Height Increment Number Height Increment Number ft, NGVD Selected ft, NGVD Selected fo9 5 9.4 1 8.0 4 9.5 1 8.1 4 9.6 1 162 3} otf 1 oS} 3 9.8 1 8.4 3 9.9 1 8.5 2 10.0 1 8.6 2 10.1 1 8.7 2 10.2 1 8.8 2 10.3 1 8.9 1 10.4 1 9.0 1 10.6 1 9.1 1 10.8 1 9.2 1 11.0 1 9.3 1 od 1 NOTE: These are the numbers of events selected for each set of 50 events. Total events selected at each height increment would be three times the numbers found in this table. area were chosen from this ranked set. This method of transferring these surge plus tide time-histories to the study area was determined during cali- bration of the storm surge model (see Part III). 22. Figure 5 shows the fourth and final part of the selection process, the assignment of probability to the selected events. The probability p , represented by each stage increment, is calculated by taking the difference of the exceedance probabilities P of the end points of the increment. If more than one event was selected to represent that stage increment, then the prob- ability assigned to that increment is divided equally among the chosen events. Table 3 contains the maximum water levels (predicted at the Boston gage) and the probabilities assigned to the three sets of selected events. The column in Table 3 labeled "Storm" refers to the numbering of the storms in Table 1. Note that since the selection process is random, not all the storms are represented in each of the three sets of 50 events (denoted as A, B, and C in Table 3), and the number of times a storm is chosen varies from set to set. In conclusion, the essence of the selection process is to choose events for simulation so that the stage-frequency curve, for a known location is 8.55 8.45 STAGE (FT, NGVD) oo ol PERCENT CHANGE OF OCCURRENCE PER YEAR Figure 5. Assigning probabilities to events selected for simulation duplicated by a limited number of events. When these events are simulated, the probability masses assigned to the events are used to construct stage- frequency curves throughout the modeled area. Table 3 Events Selected for Modeling Set A Set B Set C Max. Level Max. Level Max. Level Storm ft, NGVD Prob. ft, NGVD Prob. ft, NGVD Prob. 1 8.3 0.0223 9.9 0.0031 9.3 0.0115 10.0 0.00375 10.3 0.0018 9.5 0.0093 10.3 0.0018 10.6 0.0021 10.9 0.00115 10.4 0.0022 10.7 0.0016 -- -- 2 8.0 0.0350 -- -- 8.1 0.0250 8.3 0.0223 -- 55 a= Bits 3 7.9 0.0420 8.1 0.0250 8.3 0.0223 8.7 0.0160 9.6 0.0082 8.4 0.0183 9.8 0.0045 10.1 0.0025 8.7 0.0160 I@o1 0.0025 -- -- -- -- 4 8.2 0.0293 -- -- 7.9 0.0420 5 8.4 0.0183 7.9 0.0420 8.1 0.0250 6 9.4 0.0090 8.1 0.0250 8.0 0.0350 -- -- 8.6 0.0190 8.6 0.0190 == = -- -- 9.8 0.0045 T 8.8 0.0145 7.9 0.0420 8.2 0.0293 9.3 0.0115 -- -- 8.3 0.0223 == -- -- -- 9.7 0.0045 8 -- -- -- -- 8.1 0.0250 aa =e ae — 8.5 0.0250 ae as a2 55 8.5 0.0250 9 8.1 0.0250 8.2 0.0293 9.6 0.0082 8.9 0.0220 -- — ae Aes 10 7.9 0.0420 8.1 0.0250 8.0 0.0350 9.2 0.0135 8.4 0.0183 22 == 11 8.4 0.0183 8.8 0.0145 -- -- 8.8 0.0145 = aS ee 22 12 8.0 0.0350 7.9 0.0420 -- -- Sra 0.0160 8.8 0.0145 == == 13 8.2 0.0293 9.5 0.0093 8.1 0.0250 10.2 0.0026 -- -- -- -- (Continued) (Sheet 1 of 3) 20 Table 3 (Continued) Set A Set B Set C Max. Level Max. Level Max. Level Storm ft, NGVD Prob. ft, NGVD Prob. ft, NGVD Prob. 14 8.6 0.0190 7.9 0.0420 8.4 0.0183 8.6 0.0190 -- -- 8.6 0.0190 15 8.0 0.0350 8.1 0.0250 -- == 10.7 0.0016 9.3 0.0115 -- -- -- -- 9.7 0.0045 -- -- 16 7.9 0.0420 -- == == am 8.1 0.0250 -- =a ze = 17 -- -- 8.4 0.0153 8.8 0.0145 -- -- -- -- 8.9 0.0220 18 -- -- -- -- 8.0 0.0350 -- -- -- -- 9.0 0.0190 19 8.3 0.0223 8.2 0.0293 7.9 0.0420 9.9 0.0031 8.7 0.0160 8.2 0.0293 -- -- 8.9 0.0220 -- -- 20 -- -- 8.2 0.0293 10.1 0.0025 SS == 9.1 0.0165 10 0.0021 21 8.0 0.0350 8.0 0.0350 -- -- 8.1 0.0250 8.0 0.0350 -- -- (3)5 55) 0.0250 8.7 0.0160 -- -- 22 7.9 0.0420 8.6 0.0190 7.9 0.0420 7.9 0.0420 9.0 0.0190 7.9 0.0420 -- -- -- -- 8.8 0.0145 -- -- -- -- 9.2 0.0135 23 -- -- 8.3 0.0223 9.4 0.0090 -- -- 8.3 0.0223 -- -- a4 -- -- -- -- 8.2 0.0293 -- -- -- -- 8.7 0.0160 25 8.2 0.0293 7.9 0.0420 8.3 0.0223 8.5 0.0250 10.2 0.0026 10.2 0.0026 9.0 0.0190 10.4 0.0022 10.3 0.0018 9.1 0.0165 10.9 0.00115 10.4 0.0022 9.5 0.0093 25 25 22 QS 26 -- -- 8.0 0.0350 9.1 0.0165 -- -- 8.4 0.0183 -- -- (Continued) (Sheet 2 of 3) 21 Table 3 (Concluded) Set A Set B Max. Level Max. Level Storm ft, NGVD Prob. ft, NGVD Prob. 27 8.1 0.0250 8.0 0.0350 8.4 0.0183 8.5 0.0250 9.7 0.0045 9.8 0.0045 28 9.6 0.0082 8.3 0.0223 10.5 0.0021 8.4 0.0250 lO) 0.00035 9.2 0.0135 Viol 0.00035 9.4 0.0090 -- -- 10.0 0.00375 -- -- Vial 0.00035 22 Set C Max. Level ft, NGVD Prob. 7.9 0.0420 8.0 0.0350 9.9 0.0031 10.0 0.00275 8.4 0.0183 IOs 7 0.0016 Iol 0.00035 (Sheet 3 of 3) PART III: STORM SURGE PLUS TIDE SIMULATION 23. The WES Implicit Flooding Model (WIFM) was used as the hydrodynamic storm surge model. A detailed description will not be given in this report. The numerical and hydrodynamic features of WIFM are discussed in Butler (1978) and the application of WIFM to coastal studies is demonstrated in numerous reports (including Butler 1983). WIFM solves the vertically integrated, time- dependent, shallow-water wave equations of fluid motion using an alternating direction, implicit, finite-difference algorithm. The model allows subgrid barriers which can be non-overtoppable, overtoppable, or submerged. An impor- tant feature of WIFM is the capability for using an exponentially stretched numerical grid which permits a concentration of grid resolution in areas of interest. Also included in the code is the capability to flood or dry indivi- dual cells during a simulation. Grid Development 24. In order to model storm surge, it is usually necessary to extend the computational grid past the edge of the continental shelf and into deep water. Since it also is desirable to have small cell sizes in areas of interest, a very large number of grid cells may be necessary to model a study area uSing one grid. Consequently, in locations with a wide continental shelf, as in the present study, a two-grid system is usually developed. A global grid with coarse resolution extends throughout the study area and out past the edge of the continental shelf. A nearshore grid which extends only over the immediate study area but with much finer resolution is also devel- oped. A surge plus tide event is first simulated on the global grid. Then, using boundary conditions saved during the global run, the event is simulated on the nearshore grid. 25. The present study does not use this two-grid system. Because of the project's proximity to the NOS tidal gage in Boston Harbor, a method was devised to use the Boston tide gage in place of a global grid. Use of the Single grid resulted in considerable savings avoiding both simulation on an outer grid and stage-frequency curve generation at a connection point between two grids. This process involved setting up a single grid (Figure 6) and then calibrating the model to produce correct water levels throughout the study 23 Figure 6. Numerical grid for storm surge model area using altered Boston water levels to drive the boundary. The procedure used to alter the Boston water levels to produce the desired results is described in the section on model calibration. 26. The final grid configuration has 2,025 cells arranged in 45 rows and 45 columns. The cells with the finest resolution are 500 ft square and cover most of the areas of interest: Roughans Point, Point of Pines, the Saugus-Pines Inlet, and the initial reaches of both rivers. The cells with the coarsest resolution, located near the boundary, are approximately 1,500 by 700 ft. The grid is orientated to match the predominant direction of the river system, since the initial reaches of the rivers form nearly 90-deg angles. Wind Forcing 27. Wind speed and direction are required inputs to WIFM for the model- ing of storm surge. For this study a spatially constant but temporally vary- ing wind forcing was used. The wind data were supplied by NED using raw data from Logan International Airport and a wind data analysis computer program 24 developed by NED. The wind data were 1-min averages of both wind speed and direction reported hourly and corrected to a 33-ft elevation. The hourly wind data were interpolated to 60-sec time-steps and applied without Spatial vari- ation to the entire study area. Two factors allowed this simplified treatment of wind forcing. First, the small geographic area of the modeled area was close to the source of the wind data. Second, the use of Boston tide gage data for boundary conditions already included the effect of the wind over the continental shelf, so the local winds were needed only to locally redistribute the surge. For the 28 northeasters chosen for this study, the average maximum hourly wind speed was 33 knots and varied from 25 knots to 48 knots. The wind directions for these maximum winds varied from 0 to 292 deg (all but three were between 0 and 90 deg) The average direction of the maximum hourly values was 73 deg. Wind directions are referenced clockwise from North. Data Collection 28. During the summer of 1984, NED supervised the placement and opera- tion of five tide gages in the study area. Figure 7 shows the location of these gages. Two of the gages, Simpson's Pier and Bay Marine Lobster, were located outside the river system at Roughans Point and in Lynn Harbor, respec- tively. The other three gages (Fox Hill Drawbridge, Broad Sound Tuna, and Atlantic Lobster) are located in the Saugus-Pines River system. All of these gages were in operation from June to October 1984. No other data collection efforts were commissioned solely for numerical modeling. Bathymetric and ele- vation data for Revere Beach and throughout most of the river system were ob- tained from previous surveys conducted for beach and channel improvement proj- ects and for highway projects. Excellent data were generally available for the area east of the Salem Turnpike and for the area immediately adjacent to the abandoned highway embankment. Bathymetric data for Lynn Harbor and Broad Sound were obtained from NOS nautical charts. Model Calibration 29. Since all five study area gages were not operational during any storm and the two gages left in operation during the winter of 1985 did not experience any significant storm induced high water, the model could not be 25 FOX HILL DRAWBRIDGE NAHANT Figure 7. Location of study area tide gages calibrated or verified to a surge plus tide event. Consequently, two periods during the summer of 1985, one at spring tide and the other at neap tide, were chosen for calibration and verification of the model. 30. A 29-hour period from 0800 29 July to 1300 30 July 1984 was chosen for calibration. During this period data were available from all five study area tide gages, and both the highest tide and the largest range of the month occurred (6.7 and 13.0 ft NGVD, respectively, at Boston). Data from each gage are plotted against data from the NOS gage at Boston (Figures 8-12). The gage at Simpson's Pier went dry at -4.2 ft NGVD resulting in the horizontal lines at low tide in Figure 8. Several facts can be immediately seen from these figures. The range, phase, and MSL of the study area gages are very close to those of the Boston gage. The water levels at high tide are all within several tenths of a foot, and the phases at high tide are all within several minutes. It is interesting that Broad Sound Tuna, a river gage, has the highest tides resulting from a small upward shift in MSL. The largest dif- ferences occur at low water where the river gages show a distinctly higher and later low tide, relative to Boston. During the calibration process, adjust- ments were made in the following items so that the numerical results would 26 NGVD) STAGE (FT, NGVD) STAGE (FT, or VW &£ MN DD -1 -2 -3 -4 “3 -§ -7 C) 4 8 12 16 20 24 28 32 36 40 44 = 48 TIME (HOURS) LEGEND —__ SINPSON’S PIER ——~- BOSTON (NOS TIDE GAGE) TIDE DATA - SIMPSON’S PIER JULY 29-30, 1984 Figure 8. Tidal calibration data, Simpson's Pier 7 6 5 4 3 2 1 C) -1 -2 -3 -4 -5 -6 -7 C) 4 8 12 16 20 24 28 32 36 40 44 48 TIME (HOURS) LEGEND BAY MARINE LOBSTER ——-—- BOSTON (NOS TIDE GAGE) TIDE DATA - BAY MARINE LOBSTER JULY 29-30, 1984 Figure 9. Tidal calibration data, Bay Marine Lobster 27 NGVD) STAGE (FT, NGVD) STAGE (FT, eorerenwna YD YY -1 -2 -3 -4 -5 -6 -7 e 4 8 12 16 20 24 2 32 36 40 #44 48 TIME (HOURS) BROAD SOUND TUNA ———- BOSTON (NOS TIDE GAGE) TIDE DATA - BROAD SOUND TUNA JULY 29-30, 1984 Figure 10. Tidal calibration data, Broad Sound Tuna > 6 Ss 4 3 2 1 @ =1 -2 -3 -4 -5 -6 -7 r) 4 8 12 16 20 24 28 32 36 40 44 48 TIME (HOURS) LEGEND —_—. ATLANTIC LOBSTER ———- BOSTON (NOS TIDE GAGE) TIDE DATA - ATLANTIC LOBSTER JULY 29-30, 1984 Figure 11. Tidal calibration data, Atlantic Lobster 28 oer MU Wey DD STAGE (FT, NGVD) e 4 8 ie 16 20 24 28 32 36 40 44 48 TIME (HOURS) FOX HILL DRAUBRIDGE ———- BOSTON (NOS TIDE GAGE) TIDE DATA - FOX HILL DRAUWBRIDGE JULY 29-30, 1984 Figure 12. Tidal calibration data, Fox Hill Drawbridge closely match the tide gage data for the 29-hour period specified above. These adjustments are explained below. a. Cross-sectional areas and frictional characteristics of both channels and the bridge openings were adjusted. The minimum cell size of 500 ft was much larger than the channel width at many of the constrictions. A smaller cell size would have greatly increased modeling costs; therefore, using the 500-ft cell size necessitated that the flow be adjusted through these oversized areas by alterations in depth and friction. It would have been convenient if the depth could have been adjusted on the oversized cells so that cross-sectional areas would match between model and prototype. However, due to the large tidal range in the study area, matching cross-sectional areas would have caused the channels to dry up well above low water. Consequently, it was necessary to make these cells deeper than the area represented in the prototype would justify. Higher water levels would cause excessive flow through these oversized channels. A compromise depth was selected so that the channel would remain flowing at low water levels. At low water the opposite problem would occur. Since the depth of the channels in the study area is much greater than the compromise depth used in the model cells, the flow restriction is higher in the model than in the prototype. This causes a reduction in flow in the model at low water. Therefore, in addition to the compromise depth, frictional characteristics were made depend- ent upon depth to produce smaller Manning's n values at low 29 water and greater n values at higher water. With these adjustments, the model was able to duplicate the calibration data in the Saugus-Pines river system. In Storage in both channel and ponds in upper reaches of both rivers was adjusted in order to match elevations, particularly those measured at Atlantic Lobster and Broad Sound Tuna. Very little bathymetric data and no tidal data were available for these areas. Therefore, storage was at first estimated from USGS topographic maps and then changed during the calibration process. The final storage areas selected remained reasonable based upon the available data. As was mentioned previously, data from the Boston gage were adapted for use as boundary conditions for the model. Since the tide in the study area conforms so closely with that measured at Boston, only minor alterations to the Boston tide were necessary. The calibration process found that Boston data should be multiplied by 0.984 and shifted forward in time by 5 min before being used as boundary values. 10 31. The results of the calibration process are depicted in Figures 13- 17. These figures show excellent agreement between numerical and measured water levels during a period of large tidal range. Model Verification 32. A 32-hour period from 1000 15 August to 1800 16 August 1984 was chosen to verify the hydrodynamic model. This time period was chosen because good data were available from the five study area tide gages as well as from the NOS gage at Boston. Also, since the calibration was preformed for a spring tide, a neap tide with a lower high tide and a small range (4.8 and 8.2 ft NGVD, respectively) was chosen to verify the model. The results for the five study area gages are shown in Figures 18-22. These results show ex- cellent agreement between numerical and measured water levels for all five locations. Simulation of Event Ensemble by the Hydrodynamic Model 33. The 150 selected events were simulated on a CYBER 205 computer in three sets of 50 by the calibrated and verified storm surge model. The simu- lations of the individual events varied from 13 to 75 hours prototype time de- pending upon the number of high tides that needed to be modeled. For each of the 150 surge plus tide time-histories, all highs with still-vater levels 30 1S ) “d OO iS n ~ EBXMaeeaeeeehabaars : 5 wo < © SE SASS habees | i ‘ wb ii) ro) — SSCS E8\" Q. SSenSry = v i= Si al a mie m m m Y tu) ~ wu m oS a) = Y = a m bp ta) bad 3s fl a 2 8 rm) re z oe S ed oe PSI i S 6 | - wu 5 u My et t iS 6 =I BY) ou @ c uu a wo a (o) i wo st a oy BS Sl oe a s= are) a ry > i u =H a us uo =e o aa o < a 6 fc} vu (W] u Bi i) w | 4 ao £=z a ww it Ww u - © os aa ua -~=z aa Lond or o on u = = ee nu S fe mies Pel eg "8 ie oI & 3) ‘ -~ < , wu cy | um Ex Re Espa tae = | = | u = Ho | u - | co) Oo oO = SSS 3 Pas o ) fo) rc) ca | = - 4 ~ é Ba = © S —|] wo 4 BEE oy ese MIE < ve) a L 3) — [= =“ € a kal [os [a as a a al SP el a aca || | i a) “ a) = & ASO OESeaora as ie - 1 -e 3 4 = -6 7 CO) CON LD Si OU AO) te TT i= tee or OH TM UY B= @ WeWoOtTee Or uF WeWwote sor uF FTODdA Figure 13. Surge model calibration results, Bay Marine Lobster 31 Figure 14. MODEL ———. TIDE GAGE OE a HE ie e 13 E L E U A T I 0 N F T N G Vy D all 14 15 17 NESE ay il aa TIME, HOURS CALIBRATION - BROAD SOUND TUNA Figure 15. Surge model calibration results, Broad Sound Tuna MODEL —_——. TIDE GAGE wcoaz AN ZPOHKADCMem 13 15 17 19 el e3 25 27 e9 31 255) 35 37 TIME, HOURS CALIBRATION - ATLANTIC LOBSTER Figure 16. Surge model calibration results, Atlantic Lobster 32 TIRE, HOURS CALIBRATION - FOX HILL DRAUBRIDGE MODEL . TIDE GAGE w =“ m C3} fee Cy My CP rk GY Ce Cy) ' 1 1 ' WeWwortre eos we zeo>a a ee ees || Surge model calibration results, Fox Hill Drawbridge Figure 17. TIME, HOURS VERIFICATION - SIMPSON’S PIER MODEL ———- TIDE GAGE or onermnmmaAa Dd ' ' ' ' ' ' WewWwWotee ort ur FTOD>Q Ee ce cree | Surge model verification results, Simpson's Pier Figure 18. 33 SUG ; | : a 28 0 HOURS e TIME VERIFICATION - BAY MARINE LOBSTER MODEL ———-. TIDE GAGE A i EY Om FS & Wa Oo Tf 7G WoIwWwotTreH OFT wr FZOdAa [2 eS ee eee ee Surge model verification results, Bay Marine Lobster Figure 19. MODEL ———- TIDE GAGE ra SCL COM: Ne Se tae el eat eal onrwouwnsr ny = WIWS>tTrFrOZF LF FZODSAa TIME, HOURS VERIFICATION - BROAD SOUND TUNA Surge model verification results, Broad Sound Tuna Figure 20. 34 . TIDE GAGE MODEL VERIFICATION - ATLANTIC LOBSTER Paaeeee sae oaeia Boos Surge model verification results, Atlantic Lobster Figure 21. TIDE GAGE MODEL iets SL (eee) si peeeea eee AZIM Esc SSS rice -2 -3 -4 5 6 7 HOURS TIME, WW) °° a _ ie4 a =] Surge model verification results, Fox Hill Drawbridge Figure 22. 35 greater than 7.0 ft NGVD were included in the simulation. A constant time- step of 60 sec was used for all events. Two computer files, saving informa- tion at each of the numerical gage locations shown in Figure 23, were the main result of each simulation. The first file was a time-history of water levels at 15-min increments. This file was used both to plot the water level time-histories at each numerical gage and to provide information to the com- puter codes which calculated wave overtopping rates and interior volumes at Roughans Point. The second file listed the maximum elevation experienced at each of the numerical gages during each event. This file was used to con- struct the stage-frequency curves for the still-water locations. Both of these computer files were given to NED on magnetic tape. 36 soe [eoTUouMU Tepow azuns JO UuOTIeDOT 0 A1VOS AIEVIIVAV SI VLVG 145A41 Y3ALVM SJYSHM SNOILVDO1 YAHLO SSAYND AODNANDSAYA-JADOVLS AO SNOILVIDO1 GN3941 "€e ounstgq YALNS0 NiddOHS OT iV INVIdVAS# yr ——rrr TIHIH XO4 @ SAAYNO ININVW ‘ SNONVS ‘ ee a a ei PART IV: WAVE MODELING 34. For each event (surge plus tide), the wave climate in a 25.9- Square mile area of Broad Sound was simulated for each hour when the still- water level was above 7.0 ft NGVD. The area considered is shown in Fig- ure 24. Depths, at mean low water, range from 0 ft at the beaches to approximately 82 ft along the eastern boundary of the grid. The shallow depths in the area required the use of a shallow-water wave model. 35. A steady-state, shallow-water, directional-spectral wave model (ESCUBED) was used to perform the simulations. The required simulations actually called for the use of a time-dependent model, but the cost of using such a model was prohibitive. In lieu of a truly transient simulation, ESCUBED was run once for each hour of each event, and the resulting wave climate was taken to be representative of the conditions existing for the entire hour. 36. For each run of ESCUBED it was necessary to specify a directional spectrum at points along the eastern boundary of the grid shown in Figure 24. To do this, wave train characteristics (e.g. significant wave heights and peak spectral wave periods) were used to define the TMA spectral shape (Hughes 1984), and the resulting one-dimensional spectrum was then distributed direct- ionally. The wave train characteristics represented both sea and swell and were derived using the methods and data of WIS. 37. A total of 848 hr of simulation was made. Resulting wave heights in the lee of Nahant peninsula indicated that local wave generation in this area was inadequately simulated by ESCUBED. Hence, an additional analysis was required when winds were from the northeast. 38. Shallow-water wave growth equations were used to estimate locally generated wave heights and periods off the north seawall at Roughans Point as well as at Point of Pines and in Lynn Harbor. The total wave climate in these regions was then assumed to be a combination of these locally generated waves and the ESCUBED results. 39. It is important to note that no wave data from Broad Sound were available. Hence, it was not possible to calibrate ESCUBED or to verify its results. 40. The following sections discuss the WIS methods and data, the ESCUBED wave model, and the analysis of local wave generation in the lee of Nahant Peninsula. 38 BOSE SR ESER REED NGS Sees Ssaene oe a a a a cx 72] | a [| | EEE DLT OF SEReekezouecdaa BOOS GoonoZene aim SaaSs6 eo Ags cso cence NooNCa co 4 OSS oneoo ooo soso Bobo 30s coe a oe oo Acne Seeceoee SESUUEESSRSESE CEBU NI003; ot i aaa SOS SRERRSSRREE er eeneee DESDE REREREEV.GS20R00 BESS ocoeeeeeecogs sesogor Sooo PCSeownn -_ {BROAD BESS 0asSS50000 SN a | i i i {SOUND ASS ge See coe Ooe Ae is ec it HBOS Codooecoo Sooo soo OOSUSCCeooooeS Soeecdtacdrcatosetoctocti WIS | || TANS eT HEGSeeeS fe a | a a a | o [| eo Wf Figure 24. Numerical grid for spectral wave model 39 WIS Methods and Data 41. In late 1976 a study to produce a wave climate for US coastal waters was initiated at WES. This ongoing study, WIS, consists of three phases. Phase I (Corson et al. 1981) and Phase II (Corson et al. 1982) wave characteristics were generated by a numerical model which simultaneously propagated and transformed the waves over a discrete grid representing seg- ments of the Atlantic Ocean. Phase I acted in the deep ocean. Phase II acted over the continental shelf where, for the purpose of classifying waves, depths may be either intermediate or deep. Phase III draws upon the Phase II data to provide nearshore wave characteristics in depths as shallow as 30 ft. For all three phases, data are available at selected points referred to as stations. 42, WIS methods and data were to be used to establish the boundary con- ditions for ESCUBED. Theoretically, the ESCUBED grid could have been extended seaward as far as the nearest Phase II station (Phase II stations are approxi- mately 34 miles offshore and 34 miles apart), and the data available at this station could then have been used in the boundary conditions. The costs of computing over such a large grid would have been prohibitive. The Phase III methodology provided an inexpensive bridge between the Phase II station and the much smaller grid actually used. Phase III Methodology 43. The reader is referred to Jensen (1983) for a complete description of the Phase III methodology. A summary is given here. The Phase II results comprise directional spectra. The Phase III methodology first takes these spectra and separates them into two wave trains, swell and sea. The two are assumed to behave independently. The swell is characterized by the height H , frequency f , and propagation direction 6 of a unidirectional, mono- chromatic wave. The energy of the sea will be distributed in frequency- direction space. A one-dimensional spectrum E,(f) can be defined in terms of the directional or two-dimensional spectrum E,(f,6) which is expressed as Qn E@)a=eh wiEa(Gyeniade (1) 0 4O 44, The Phase III methodology assumes that, at the Phase II station, E,(f) can be represented parametrically using only two parameters: the energy based significant wave height Ano and the frequency of the spec- tral peak f,, . This one-dimensional spectrum is then given a directional distribution using the following equation: BGG) = Bah S= case WO = A) (2) aoe 1 3 m Here, on is the central angle of the spectrum. E,(f,@) is discretized so that each component can be propagated from the Phase II station to the Phase III station in accordance with linear wave theory. 45. The Phase III methodology assumes straight and parallel bottom contours so that refraction and shoaling of swell and of the discrete ele- ments of E,(f, 8) may be determined analytically. The sea is further trans- formed by wave-wave interactions. Depth-controlled criteria limit both 4H and Ano . Sheltering by capes or peninsulas is included in the Phase III methodology. 46. Refraction, shoaling, and depth limitation acting on the swell transform H, ff, and 6 at the Phase II station into new values in shallow water at the Phase III station. If the Phase III station were sheltered from the swell, then H is zero. Refraction, shoaling, wave-wave interactions, and sheltering acting on individual components of the sea result in a new spectrum for sea at the Phase III station. Ano 5 ie , and Q). are extracted from this spectrum. 47. The final Phase III result comprises six wave characteristics: H , f , and 6 of the swell and Ano » f, » and On of the sea. The wave climate at the Phase III station is taken to be completely defined by these six parameters. Use of Phase III Methodology for Broad Sound Wave Climate Simulations 48. WIS Phase II, sta 13, directional spectra were used as deepwater input. This station is located at latitude 42° 32.5' N and longitude 70° 14' W. The Phase III station was positioned at latitude 42° 23.5' N and longitude 44 70° 53.5' W (Figure 24). This puts the Phase III station approximately 4.6 mi due east of Roughans Point in 75 ft of water. Cape Ann, to the northeast, and Cape Cod, to the southeast, provided some shelter for the Phase III station. The sheltering was such that only those waves approaching from between N40° E and S60° E could reach the Phase III station. 49. Phase III results were produced at 3-hour intervals. Linear interpolation was used to calculate H, f, 6,H ; fn , and Ue for every hour. Wave Climate Simulations for Broad Sound Summary of ESCUBED 50. The reader is referred to Hubertz (1985) for a detailed discussion of ESCUBED. Relevant aspects of the model are presented here. 51. Essentially, ESCUBED propagates components of discrete directional spectra over a user specified bathymetry. Calculations proceed to propagate individual components of these spectra across a rectangular, uniformly spaced finite difference grid. 52. The grid used for the wave climate simulations at Broad Sound is shown in Figure 24. The grid spacing in both the x and y directions is 656 ft (200 m). At each grid point, the energy of the individual components of a spectrum is limited by the finite depth water equilibrium range proposed by Kitaigorodskii, Krasitskii, and Zaslavakii (1975). The range, which applies for frequencies greater than the peak, is a function of depth and frequency. This limitation could be thought of as an energy sink where the energy loss is through turbulent and viscous processes associated with white capping and large scale breaking. Determination of a spectrum for the ESCUBED boundary condition 53. The Phase III wave characteristics for sea, H, , and f, were used to generate a TMA spectrum Epy,(f,h) (Hughes 1984). The TMA spectrum is representative of fully developed wind seas in finite depth water. 54. The TMA spectrum was evaluated using the depth at the Phase III station, i.e. h = 75 ft . Let Epya(f,75 ft) = Epya(f) . The one- dimensional spectrum Enma(t) was distributed directionally using a cost (6 - a) spreading. No energy was allowed to have a direction outside 42 the WIS Phase III sheltering angles. The total energy of the one-dimensional and directional spectra must be equal. This requirement is expressed as i Sele = J os Esea(f 9) de df (3) where E seq (fs 8) is the directional spectrum of the sea along the eastern boundary of the ESCUBED grid. 55. Assuming the relationship shown in Equation 4, Equation 5 can be derived from Equation 3. The « in Equation 5 is a constant which is determined by Equation 6. The limits of integration in Equation 6 match the Phase II sheltering angles since the energy density outside these angles is zero, as indicated below. Esea(fr®) = « we (Do 0.) Enya(f) (4) o oT mM co JP BAC) Gee on Kk cos’ (@ - 6) do Z Bayne) ate (5) =i (5/18) x a ws () S o-) de (6) (-1/6) 1 The continuous spectrum, E seq lf) is discretized using a frequency incre- ment Af = 0.01 Hz anda direction increment Ae = 20 deg . Let E_(f.,0.) be this discrete spectrum. sear deal 56. The final step in determining a directional spectrum representing both sea and swell for the boundary condition at the eastern side of the grid is to add the swell to E (f;,8;) . The swell can also be represented by a sea discrete spectrum, E swell ti? i?) . If the energy of the swell is uniformly distributed over one frequency-direction band of the spectrum, then a discrete directional spectrum E (f,,8,) can be written as follows: swell 43 all ail e2 ee fo an Sy and S, < S, (14) where Sw = height of wall section Sp = ocean water level Nadewe 1.5 0.385 a 1.5 fe) W ; Qout = Co(2.7)(S; - S.) 1 - segue sie S; > S, and So > S, (15) it W Qout = &6 (16) 75. The coefficients which were used in the above equations are shown in Table 8. Coefficients Cy, and Co are the length of weir section. Coef- ficient Cg is the pumping rate in cubic feet per second. The increase in 73 for the proposed condition is due to improved inlet design. Coefficient C3 (see page and drainage) increases for the proposed condition because of the addition of a gravity drain (see Figure 33). Note that two values are listed for coefficients C4 and C5 . For existing conditions, reach D was divided into two sections for this analysis, one section at a height of 60 Table 8 Coefficients Used in Outflow Equations 13-16 Coefficient Existing Proposed C3 yy 94 Cy 250 Sree 5715 ae € 250 =o ? 515 oe Ce 38 50 10.5 ft NGVD and the other section at a height of 11.5 ft NGVD (the first and second numbers, respectively). The overflows at the west end of the Roughans Point area are included only in the 11.5-ft coefficient. 76. In the flood routing calculations, the path and time of travel of the water from the time it overtopped the walls until it reached drainage points were not considered. Therefore, all water entering Roughans Point was assumed to be immediately available for drainage. The characteristics of inlets and the capacity of the system were taken into account in the coeffi- cients of Equations 13-16. The flood routing calculations can be summarized as follows. A 1-minute time-step was used. Inflow volumes from wave over- topping from all reaches were combined and then added to the volume remaining from the previous time-step. Outflows were subtracted using the methods out- lined in the previous paragraphs. For each time-step the resulting stage was determined from a stage-volume relationship supplied by NED (Figure 34). Finally, the maximum stage during each event was determined for use in stage- frequency generation. 77. Sufficient data were available during two historical events, Feb- ruary 1972 and February 1978, for a rough calibration and verification of the combined overtopping and flood routing process. The maximum interior flood level which occurred during these two events was estimated from water marks and eyewitness accounts. For calibration of the Roughans Point interior cal- culations, the 1978 event was simulated by the storm surge and wave models. Then the overtopping rates and maximum stage were calculated by the computer codes described above. The first attempt predicted interior stages which were in excess of those observed; therefore, refinements and adjustments, discussed 61 STAGE (FT, NGVD) sie te Se es ed \) 5® 100 150 200 250 300 358 VOLUME (ACRE-FT) Figure 34. Stage volume versus volume relationship for Roughans Point interior in the following paragraphs, were made to match the estimated stages. 78. For existing conditions, overtopping at reach D was not allowed during periods of weir outflow at that wall. Since there would be a contin- uous current flowing outward in this situation, it was reasoned that any over- topping would almost immediately be conveyed back into the ocean. Without this reasoning, reach D would contribute enormous quantities of overtopping at those times when Roughans Point was full to overflowing. This assumption is consistent with the limited information available from the only historic event, in February 1978, during which the water level inside Roughans Point was higher than the elevation of reach D. 79. Wave heights attacking reach E were reduced by 15 percent. There were several possible adjustments which could have been made to eliminate overprediction of overtopping rates. Among these are (a) reducing the cal- culated overtopping rate, (b) lowering the still-water level, and (c) reducing the wave heights. The wave heights were selected for reduction because they are the least certain of these possibilities (see "Estimating Error in Stage- Frequency Curves" in Part VI). 80. Wave heights were also lowered for the three northern reaches. At 62 reach D , the height of waves which propagate from the open ocean was set to zero. There were two justifications for this adjustment. First, due to the orientation of the wall, there is no opportunity for these waves to attack the wall from any but very oblique angles. Second, reach D would be partially sheltered from waves from these oblique angles by the tip of Roughans Point and by Simpson's Pier. At reaches A and C, waves from the open ocean were re- uced by 50 percent. As at reach D, these waves would approach from an oblique angle; however, refraction would turn these waves more normal to reaches A and C than at reach D. Since the physical modeling assumed a wave direction nor- mal to the structure, using the full wave height for these waves would result in the overprediction of overtopping rates. The locally generated waves were reduced by 15 percent for all three north wall sections. This can also be justified by the fact that these waves do not always approach normal to the wall sections. 81. The above adjustments were made to the overtopping calculation and flood routing computer codes to match calculated values of interior stage to those observed during the February 1978 storm. The February 1972 storm was then simulated to verify the revised procedure. The results of these two simulations are compared to estimates of actual flooding in Table 9. Table 9 Comparison of Calculated to Observed Flood Stage Calculated Observed Storm ft, NGVD ft, NGVD 1978 1169) 11.8-12.0 1972 9.6 8.8-9.0 82. The results of this calibration and verification were judged to be acceptable. The 0.6-ft difference between observed and calculated water levels for the 1972 storm seems reasonable when considering that the calcula- tions were based upon a stage-volume relationship determined from 2-ft contour intervals. Simulation of the Event Ensemble by the Flood Routing Model 83. Following calibration and verification of the flood routing model, 63 events were simulated for the existing one and five alternative structure com- binations. Inputs to the model were the time-histories of overtopping rates for each of four Roughans Point reaches (A, C, D, and E). Six different com- binations of the northern and eastern reach structures were modeled. Since the north wall has only two structure classes, "Existing" and "Original Pro- posal," a combination of northern and eastern structures was given the name of the eastern structure. The "Existing" combination is self-explanatory. The "Original Proposal" combination is made up of the northern and eastern structures proposed before the beginning of modeling (see Figures 28-31 and NED 1983). The other four alternatives, "Wide Berm," "Two Berms," "Wide Berm + 1-ft Cap," and "Wide Berm + 2-ft Cap," combined the eastern structure of the same name (see Figure 32) with the northern "Original Proposal" structure. There were two output files. One file was a time-history of flood stages for each event and structure simulated. The second file contained the maximum stage during each event for each combination of structures simulated. This second file was used to compute the stage-frequency curves. North Wall Tests 84. During the course of simulating overtopping and flood routing, there was no contribution to overtopping volumes from the "Original Proposal" northern structure. Tests were conducted to determine the effect of lowering the height of the protection along the whole north side of Roughans Point. Since no additional physical model tests were to be run, a method had to be devised to use physical model data from the proposed northern structure (17 ft). Reconsideration of the overtopping rate equation (Equation 8), re- veals that changing the height of the northern structure would only change one term in that equation, namely F , the freeboard. Since the water level would not be changed, the characteristics of the waves attacking the structure would not be changed. Therefore, even though lower heights were not tested, esti- mates of the overtopping for lowered structure heights could be made by re- ducing the freeboard in Equation 8. Using the February 1978 historic event for the initial tests, the northern structure was lowered in 1-ft increments. For this event, overtopping did not start until the structure was lowered to 14 ft NGVD, and large volumes of overtopping did not commence until a struc- ture height of 12 ft NGVD was tested. Using these results, the full ensemble 64 of 150 events was simulated for northern structure heights of 14, 13, and 12 ft. The results of these tests are presented in Part VII. 65 PART VI: STAGE-FREQUENCY CURVES 85. In this section, the method for establishing stage-frequency curves will be described for both the still-water locations and for the interior of Roughans Point. 86. The goal of this project was to produce stage-frequency curves for two distinct processes. The first process involved the interaction of storm surge and tide to produce still-water levels at coastal (and river) locations, and the second process combined waves with the surge and tide to produce flood levels behind seawalls due to wave overtopping. Although the simulation of these two processes involved some different steps, development of frequency curves for the two processes once the water levels are determined is essen- tially the same. 87. Probability was assigned to each of the events selected for simula- tion, as described in Part II. By assigning the probability to the maximum still-water level caused by the event at each numerical gage location, stage- frequency curves can be constructed by the following method. First, an array of possible stages at each gage location is established with a discretization interval (0.1 ft for this project). Next, for all 150 events, the probability masses assigned to each event are accumulated in the stage interval which brackets the maximum water level that occurred for that event. Exceedances can be determined for any interval by adding the probability of that interval to the sum of the probabilities of the intervals above it. After this was accomplished for the total set of 150 events, the process was repeated for each of the three sets of 50 events. This produced three additional sets of stage versus exceedance relationships which were used for confidence calculations. 88. The range of stages modeled in the still-water level portion of this study was just over 3 ft (from 7.9 to 11.2 ft NGVD). All of the re- sulting 33 discretization intervals did not receive probability. Some intervals received probabilities from several events causing in places (in the array of stages) a series of heights where no event deposited its probability. This occurrence results in a jagged line when the stage-frequency is plotted. 89. There is no physical reason why adjacent height intervals should have greatly different probabilities. The jagged nature of the raw curves is caused by trying to represent a continuous process (all possible storm events) 66 with a discrete process (50 storm events). Modeling more events would result not only in a smoother curve but also in greater expense. For example, had 500 events been modeled, it would be highly unlikely that one height interval would receive the probabilities of several events while the three intervals below received none. Therefore, if an economically feasible number of events were to be modeled, the raw output of the stage-frequency generation would require smoothing to adequately represent a continuous curve. 90. Smoothing was accomplished using linear regression of the stage- frequency data when plotted on an appropriate probability paper. Equation 17 is a formula for the construction of Weibull probability paper. Where a c Xnew =[-1" (Xo1a)] (17) the variable xX ;q is the inverse return period, x,,, is the transformed abscissa value, and ec is the variable to be adjusted to best represent the data with a straight line. After numerous trials a ec value of 0.80 was chosen. Figure 35 contains a plot of both the raw and regressed stage- frequency curves for the Fox Hill Drawbridge still-water location. [=] > o Ze i w uJ o xc = wn HK RETURN PERIOD, YRS Hat STAGE-FREQUENCY — — —— REGRESSED STILL WATER LEVEL REGRESSION CHECK FOX HILL DRAWBRIDGE Figure 35. Example of raw and regressed stage-frequency curve 67 PART VII: RESULTS AND CONCLUSIONS Roughans Point 91. The stage-frequency curves for the interior flood levels in Roughans Point are presented in Figures 36-40. For these curves it was not possible to regress the total curve as was explained in Part VI. The physics of the problem undergoes a sudden change at higher levels where the effect of weir outflow limits the capacity of the interior of Roughans Point. Also the extreme lower portion of the curves does not conform to the straight line tendency of the middle portion. The lowest possible stage is 3.6 ft NGVD corresponding to the lowest point inside Roughans Point. Consequently, the stage-frequency curves remain at 3.6 ft until the onset of overtopping. Therefore, the linear regression was limited to the middle segment of each curve for the Roughans Point stage-frequency curves. Smoothing for both the lower and higher segments of the curves was done by eye. 92. As explained in paragraph 83, flood levels were calculated for six different combinations of northern and eastern structures. The names of the structure combinations plotted in Figures 36-40 refer to the names of the eastern component. (For the "Existing" and "Original Proposal" structures (NED 1983) refer to Figures 28-31, and for the other alternatives refer to Figure 32.) Three tests lowering the height of the proposed north wall struc- ture were conducted. Since it was determined that there was a negligible difference between the curves with the originally proposed north wall height (17 ft NGVD) and the curves from the highest of the three additional tests (14 ft NGVD), curves resulting from the 17-ft height are not presented. Curves for the six structure combinations are shown in Figures 36-38, with the height of the northern structure in the three figures being 14, 13, and 12 ft, respectively. Note that the 14-, 13-, and 12-ft north structure heights refer only to the alternative structure combinations. For the "Existing" curve, shown on these graphs for comparison purposes, the northern structure is set at the existing height for each structure section. 93. Using Figure 36, several features of the stage-frequency curves for the interior flood levels at Roughans Point will be discussed. The greatest differences among the alternatives occur at the lower return periods. Near the 500-year return period all six curves tend to come together. As was 68 EE NIE Oo = Oo Zé = ire uJ o a = wn \ \ \ \ , . , a, : . , mS WERE IN \ , ae ; BY ual, ray RETURN PERIOD, YRS EXISTING STAGE FREQUENCY — — — — ORIGINAL PROPOSAL ROUGHANS POINT INTERIOR PE UarEne NORTH UALL - 14.0 FT ——--—-UVIDE BERN + 1 FT CAP —— --—_VIDE BERM + 2 FT CAP Figure 36. Roughans Point stage frequency, northern structure height = 14 ft STAGE (FT NGVD) | GEBE CBRGRR Ze JOSUIAS GSS HANH RETURN PERIOD, YRS LEGEND EXISTING STAGE FREQUENCY — —_____ ORIGINAL PROPOSAL VIDE BERN ROUGHANS POINT INTERIOR TUO BERNS = ——..—.UVIDE BEKn + 1 FT CAP NORTH WALL 13-0 FT ——_---—_UIDE BERM + 2 FT CAP Figure 37. Roughans Point stage frequency, northern structure height = 13 ft 69 {@) > Oo Zz b w Ww oO a = wn RETURN PERIOD, YRS ESE STAGE FREQUENCY _— — —— ORIGINAL PROPOSAL ROUGHANS POINT INTERIOR Tuo BERAS ——..—.UIDE BERN + 1 FT CAP NORTH WALL - ie.@ FT —__--—_UIDE BERM + 2 FT CaP Figure 38. Roughans Point stage frequency, northern structure height = 12 ft STAGE (FT NGVD) TEEPE TTT RT ELUTE, RETURN PERIOD, YRS STAGE FREQUENCY ROUGHANS POINT INTERIOR NORTH WALL COMPARISONS WIDE BERM Figure 39. Effect of northern structure height on the "Wide Berm" alternative 70 STAGE (FT NGVD) ST 10 RETURN PERIOD, YRS STAGE FREQUENCY ROUGHANS POINT INTERIOR NORTH WALL COMPARISONS WIDE BERM + 1 FT CAP Figure 40. Effect of northern structure height on the "Wide Berm + 1-ft Cap" alternative discussed in Part V, Roughans Point has a limited volume capacity. When the amount of overtopping surpasses the capacity of the interior, the water pours out over roadways into another drainage area. Therefore, although the various alternatives are still producing very different overtopping rates, the flood levels that result are similar for the highest return periods. 94. Although the "Two Berms" alternative produced results very sim- ilar to the "Wide Berm + 1-ft Cap" alternative, the "Two Berms" structure is not a recommended alternative. The physical model tests showed the "Two Berms" structure was not stable. For details see Ahrens and Heimbaugh (in preparation). 95. The wide berm configuration proved to be effective in lowering overtopping at the still-water levels which accompany return periods less than 100 years. Notice, however, that in Figure 36 the "Wide Berm" curve crosses above the "Original Proposal" curve at about 150 years. The berm loses its effectiveness in reducing overtopping as the higher still-water levels sub- merge it. 96. The effectiveness of the berm is dramatically improved by adding 71 height to the wall behind it, as is seen in both the "Wide Berm + 1-ft Cap" and the "Wide Berm + 2-ft Cap" alternatives. Studying the overtopping rate equation (Equation 8) shows that the overtopping relationships developed from the physical model are very sensitive to freeboard and, therefore, to struc- ture height. 97. Recommending a height for the north wall is difficult. None of the three heights were actually modeled by the physical model. The final struc- ture selected must, of course, result from a detailed economic analysis. The technique of using the 17-ft north wall physical model results to predict the results for lower revetment heights by lowering freeboard was the best avail- able but must lower confidence in the analysis. The choice seems to be be- tween the 13- and 14-ft heights. The 12-ft height allows significantly great- er overtopping to occur. Figures 39 and 40 show the effect of north wall height on the "Wide Berm" and the "Wide Berm + 1-ft Cap", respectively. Since the height of the existing wall sections at A and C (15.3 and 13.7 ft NGVD) is higher than that of the 13-ft trial, the best choice would be a revetment at a 13-ft height with the wall keeping its existing height at A and C, with the height at B being a transition between A and C, and the height at D matching that at C. Still-Water Locations 98. Stage-frequency curves for 14 locations within the Saugus-Pines River system and the coastal areas bordering Broad Sound are presented in Figures 41-54. Figure 23 shows the location of these 14 numerical gages. Just prior to the completion of the study, additional data were collected by NED during the highest predicted tides of September, October, November, and December 1985 for several locations in the extreme upriver portions of the modeling area (Figure 55). Because of increased interest in flood protection for these areas, it was hoped that the additional data would allow adjustment of the modeling results upstream of where calibration data were previously available. Data were collected also at the Fox Hill Drawbridge calibration gage location, and data for the Boston tide gage were obtained from NOS. These data are summarized in Table 10. 99. Based on the information shown in Table 9, adjustments were made to those numerical gage locations west of the abandoned highway embankment and T2 (GADN 14) 3961S YRS RETURN PERIOD, STAGE-FREQUENCY STILL WATER LEVEL SIMPSON’S PIER Still-water level stage-frequency curve, Simpson's Pier Figure 41. STAGE-FREQUENCY STILL WATER LEVEL POINT OF PINES a (=) ° - a WW a z= [4 =) = re) x Bee AERE BBs eee U eC EEBBE: SO a So co et StH aL J en CC eC we (QADSN 14) 3981S Still-water level stage-frequency curve, Point of Pines Figure 42, 1 ALTTTTE TEAC TATA TTT, IM ALT GREE MEG ELIGbIl: ene ere EEE Ves Sao ©) mcs Gamo maa (QAON L4) J9b1S . WRG 6 a (=) - a uw a z a =) te WW a STAGE-FREQUENCY STILL WATER LEVEL BAY MARINE LOBSTER Still-water level stage-frequency curve, Bay Marine Lobster Figure 43. STAGE-FREQUENCY REVERE BEACH STILL WATER LEVEL YRS RETURN PERIOD, S is) CGE MT ERIE RBIER oe EERE (GAIN 14) 3981S Still-water level stage-frequency curve, Revere Beach Figure 44, 74 TTT EAE (QASN 14) J9KLS YRS s a o _ "4 yy a z= a 2 e uJ a STAGE-FREQUENCY STILL WATER LEVEL RIVERSIDE Still-water level stage-frequency curve, Riverside Figure 45. (QAIN 14) 3981S YRS S a ° - a ey) a z 4 =) = Wy e STAGE-FREQUENCY STILL WATER LEVEL GE PLANT Still-water level stage-frequency curve, GE Plant Figure 46. (2 oe i CNT He EBESoSei JOC TE (QAON 14) 3981S ” vps S a fo) ~ [e4 WwW Qa z ie4 2 = ee) ac STAGE-FREQUENCY STILL WATER LEVEL BROAD SOUND TUNA Still-water level stage-frequency curve, Broad Sound Tuna Figure 47. ALTE TATE TERA IATLMHEEREAD CAE STAGE-FREQUENCY STILL WATER LEVEL OAK ISLAND YRS RETURN PERIOD, TTT NTH Bi elaeisis e/le/6/- HUE NOE SESE | oY = Sg or on won a =a on (QAIN 14) JOULS Still-water level stage-frequency curve, Oak Island Figure 48. 76 EEEEPET TTT (QAON 14) JDKLS YRS a (=) ° io 4 uJ a z= a =) = WW a STAGE-FREQUENCY STILL WATER LEVEL FOX HILL DRAWBRIDGE Still-water level stage-frequency curve, Fox Hill Drawbridge Figure 49. STAGE-FREQUENCY STILL WATER LEVEL ATLANTIC LOBSTER a a °o Lan] a W a z ie 4 =] = WW [e4 IN PEOEEBEEEEEEEE COREE ENERGIE: (GASN L4) 3981S Still-water level stage-frequency curve, Atlantic Lobster Figure 50. tat AIT ATTA TERETE? BEGET ECO BEG ae wy WS Oy OS & OD Ct BS 2 (GAIN 14) 3591S YRS Ss (=) (=) i) (4 uJ a z= a = te WW a STAGE-FREQUENCY STILL WATER LEVEL SHOPPING CENTER Still-water level stage-frequency curve, Shopping Center Figure 51. (QAQN 14) 3981S YRS RETURN PERIOD, STAGE-FREQUENCY STILL WATER LEVEL MODEL - AFTER ADJUSTMENT SEAPLANE BASIN Still-water level stage-frequency curve, Seaplane Basin Figure 52. 78 it | () > (S) Zz = aL = uJ o Cc — w RETURN PERIOD, YRS LEGEND STAGE-FREQUENCY MODEL _ AFTER ADJUSTMENT STILL WATER LEVEL SAUGUS CURVES Figure 53. Still-water level stage-frequency curve, Saugus Curves _——— a ie rie STAGE (FT NGVD) RETURN PERIOD, YRS ea oe STAGE-FREQUENCY ———. AFTER ADJUSTMENT STILL WATER LEVEL UPPER SAUGUS Figure 54. Still-water level stage-frequency curve, Upper Saugus 19 2 O E 2) O ca AVENUE BRIDGE gin 80 wt 4 Locations of additional water level data collection Figure 55. Table 10 Maximum Tide Elevation* Data for Upstream Areas Date = Location 9-17-85 10-15-85 11-13-85 12-12-85 Boston 6.19 7.44 1220 8.01 Fox Hill Drawbridge 6.55 7.65 fe 15 S61 (ese) Boston Ave. Bridge 6.1 TBD 6.95 7.9 Town Line Brook 6.1 Uo 6.65 -- East Saugus -- -- -- 6 (est.) # All elevations are for the maximum elevation and are in feet referenced to NGVD. upstream of the Fox Hill Draw Bridge. For areas west of the embankment, the curves were lowered by 0.5 ft at 1.5 years, and a straight line was drawn between this level and the original curve at 200 years. The same adjust- ment was made for the areas west of the drawbridge, except the reduction at 1.5 years was 0.3 ft. The adjustments were phased out at the higher water levels because, as the water level increases, the access of floodwaters to the back locations in the study area improves. The effect of storm surge is to raise the sea level upon which the tide propagates. Unlike a hurrricane surge time-history, which can be sharply peaked because of rapid changes in both speed and direction of the winds, the time-history of almost all the north- easter surges is very broad with any rapid fluctuations in water level confined to several tenths of a foot. At these higher water levels channel cross sections are increased, the effect of bottom friction is lessened because of greater depths, and new paths of access are created from the overtopping of barriers (roadways). All these factors would tend to negate losses seen in Table 9 for higher water levels. The adjustments to the above mentioned areas are shown as the dotted lines in Figures 52-54. 100. Differences among the curves presented in Figures 41-54 are small. This small difference is not surprising considering the small size of the area being modeled. In general, the curves are slightly higher for locations in- side the Saugus-Pines River system as compared to locations in Broad Sound. The predominant wind directions during severe northeasters are from the north- east to north. On the inside of the inlet, these directions would tend to 81 push water up the Pines River away from the inlet, pumping more water into the river system. Curves for locations upstream on the Pines River would be fur- ther increased by the effect of the wind setting up the water over the shallow marsh areas. In general there was a small north to south gradient in flood levels with the more southern areas higher by one-half to three-fourths of a foot during the more severe events. For the Broad Sound locations a smaller variation of a few tenths of a foot with the higher levels at the more south- ern locations also is explained by the direction of the winds. 101. Stage-frequency curves are not presented for the marsh areas west of the highway embankment. Modeling the routing of the floodwaters in these areas is beyond the scope of this study. For lower return periods observa- tions indicate there is a head loss as the waters go north from the Pines River channel across the Saugus Marsh. In these areas, at lower return periods, flow is contained in drainage ditches which are too small to model with the present grid resolution. Also, other subgrid effects such as loca- lized areas of high ground which could thwart the movement of floodwaters are important but were not considered. 102. It is important to emphasize that the effects of ice and snow were not taken into account by the storm surge modeling. It is possible and perhaps even likely that severe northeasters would be accompanied by heavy accumulations of snow and ice formation in the river systems. Snow banks formed from the clearing of roadways could act to divert floodwaters and provide some measure of protection to some areas. Ice could restrict bridge and channel openings and, therefore, reduce the amount of water entering back areas. Ice cover of open water would likely reduce the wind setup of the marsh areas. Although the above mentioned effects indicated the effect of ice and snow would be to reduce flood levels, scenarios are possible where the opposite would be true. For example, ice could divert the flood into areas which would not have been affected without the diversion. Estimating Error in the Frequency Curves 103. The final products of this study are curves which depict stage versus return period for flood levels at many locations throughout the study area. At any one return period, say 100 years, the curve is merely an esti- mate of the true flood level. Moreover, this estimate is only a point 82 estimate which represents a random variable which has a probability distribu- tion. If this probability distribution can be determined, confidence intervals could be calculated by specifying the probability that the true flood level lies between a range of heights about the estimated value. Confidence inter- vals are relatively easy to determine when dealing with a single data set, for example, confidence intervals about the mean value of a set of data. However, the calculation of stage-frequeney curves as done in this study involves mul- tiple data sets and multiple modeling systems. Even if it were possible to determine confidence intervals about each of the processes separately, there would still be the problem of combining separate intervals into one interval for the final stage-frequency curve. The total 90 percent confidence interval would not be the sum of the 90 percent confidence intervals of all the pro- cesses. For example, the storm surge model may overpredict, the wave model underpredict, and the probability model assign too low a probability. Conse- quently, no attempt will be made to place error bounds on the final curves. Instead, a verbal description of the types and, where possible, the magnitudes of the various sources for error will be given. A method has been developed to show curves for the error associated with the process of selecting a limited number of events to be modeled from the infinite number of possible events. Since the physical modeling was not a part of this report, no attempt Will be made to determine the potential for error from the physical model- ing. The reader will have to analyze the following paragraphs and determine how the possible error will influence any engineering decisions. 104. The modeling of still-water level involved three main parts: data collection and analysis, numerical model calibration, and simulation. The tide gage data used in the project were carefully screened to remove spurious data points; therefore, this information was probably corrected to about 0.1 ft. Calculating accurate tide time-histories was difficult. Five sets of tidal constituents, each based on an analysis done for a different time period, were tested. Due to the large tidal range at Boston, slight errors in the phase of the predicted tide can cause significant errors when calculating the storm surge time-histories. The storm surge time-histories used for combination with tide were edited by eye to remove any errors caused by poor tide prediction. The numerical grid, as shown by the calibration results, had sufficient resolution to accurately model tide in the areas where calibration data were available. WIFM has performed well in numerous studies, and the 83 calibration and verification in this study produced excellent results. The one-grid system used in this project should prove to be much more accurate than a two-grid system because of the lack of wind data needed to force an outer grid. The major source of potential error in the water level modeling is the lack of storm data for calibration of the model. Implicit in cali- brating the model to tide alone is the assumption that the magnitude of the storm surge at the Boston tide gage is very close to the magnitude of the storm surge in the study area for any storm event. Because the two locations are so close to one another in comparison to the size of either the continen- tal shelf or the size of a typical northeaster, this assumption is probably more accurate than the alternative of using a two-grid system. Taking all these factors into account, it is estimated that the accuracy of any one simu- lation of the storm surge model would be within a few tenths of a foot in areas close to the tide gages and within about one-half foot in those areas west of the highway embankment. 105. The wave modeling portion of the project was less accurate than the water level modeling for four main reasons. First, the state of the art in wave modeling, particularly in shallow water, is not as advanced as in surge modeling. Second, the numerical wave model used is more recent than WIFM and, therefore, less well tested. Third, there were no wave data available for either calibration or verification of the model. And, fourth, the boundary conditions for the wave modeling (the WIS hindcasts which are the best available) were not as accurate as the gage data used for the water level modeling. These four factors are somewhat offset by the fact, that, for all the more severe wave conditions and for many of the times when overtopping occurred at Roughans Point, the waves approaching the wall were depth limited. 106. The flood routing model contained a series of assumptions for cal- culating outflow from the interior of Roughans Point. For the flood levels bracketed by the 1972 and 1978 floods, the flood routing model should produce good results. However, for extreme floods, the interior water level is heav- ily dependent upon the volume of water leaving the interior by flowing over roadways and, for existing conditions, over reach D. Therefore, flood levels higher than those produced by the 1978 event are more uncertain than are lower flood levels. 107. The probability model contained several processes which could potentially introduce error into the final curves. These included assigning 84 probability from the NED Boston stage-frequency curve, selecting events to model, and fitting a curve to the raw modeling results. 108. It is beyond the scope of this report to assign error bounds to the NED stage-frequency curve. However, a simple investigation of the pos- sible error in the curve would be as follows. The curve was based upon 131 years of record, 57 of which were from a continuous record at the NOS tide gage. Because of the relatively long record, the bottom portion of the curve (i.e. return periods of less than 15 years) should be very accurate. The middle portion of the curve (i.e. return periods between 15 and 100 years) is within the length of record and should be accurate to within a few tenths of a foot. The portion of the curve above the 100-year return period would be more uncertain with, of course, the uncertainty increasing with return period. However, because of the extremely flat nature of the curve (there is only a 1-1/4 ft difference between the 50- and 500-year levels), it seems safe to predict that the curve should be accurate to within a half foot even at the 500-year return period. 109. The potential error from the curve fitting process can be best seen in plots of raw versus regressed output. For Fox Hill Drawbridge, the raw and the regressed still-water level stage-frequency curves, previously presented in Figure 35, had a linear regression correlation coefficient of r = 0.997 . Figure 56 shows the raw versus regressed output for the "Wide Berm + 1-ft Cap" alternative at Roughans Point which had a correlation coefficient of r=0.994 . These correlation coefficients are representative of those occurring at all locations. The regression was highly accurate and poten- tially introduced only minor error into the total process. The lowest corre- lation coefficient was greater than 0.98 for both the still-water level loca- tions and the interior of Roughans Point. Determining Error Bands for the Selection Process 110. The selection process that determined which events were selected for modeling was designed specifically for this project. As a result of limited experience with this technique, it is much more difficult to determine the potential error of the selection process as compared to the potential error of the more familiar processes of data collection, data analysis, and numerical modeling. In order to estimate the variability of the selection 85 CUE STAGE (FT NGVD) TATE i EA 8 ea La tae Gl RES bua ies Ral ai] ok ea eae) iain | RETURN PERIOD, YRS STAGE FREQUENCY ROUGHANS POINT INTERIOR REGRESSION CHECK UIDE BERM + iFT CAP RAW — — — — REGRESSED Figure 56. Raw and regressed stage-frequency curves, "Wide Berm + 1-ft Cap" process, the 150 events were divided into three sets of 50 events each. Each of these sets was processed separately producing three stage-frequency curves. These three curves were generated for each numerical gage for the still-water level locations as well as for each of the six structure combinations at Roughans Point. As was mentioned in Part II, 150 events were more than necessary to produce consistent results for the still-water locations. This assertion was confirmed when stage-frequency curves derived separately from the three sets of 50 events were plotted for each still-water location. For most of the locations there was not a discernible difference between curves from the three sets. Figure 57 contains the three stage-frequency curves for Oak Island which had the greatest variation of all the still-water locations. As can be seen from this figure the variation resulting from selecting a limited number of events to represent all possible events is negligible for the still-water locations. 111. The potential error caused by the selection process is much greater for stage-frequency curves for the interior of Roughans Point than for the stage-frequency curves for the still-water level portion of this project. The flooding levels in the interior of Roughans Point are dependent not only upon 86 Oo > o Ze Ll re LJ ao c = w RETURN PERIOD, YRS STAGE FREQUENCY STILL-WATER LEVEL THREE SET COMPARISON OAK ISLAND Figure 57. Stage-frequency curves, three-set comparison, Oak Island the still-water level of the event but also upon the magnitude, direction, and duration of the event's waves. Since the selection process used still-water level as its only criterion, two events which were selected with the same still-water level could have very different wave characteristics and, there- fore, could cause very different flood levels at Roughans Point. Stage- frequency curves for the three sets of 50 events are shown in Figures 58 and 59 for the "Wide Berm" and the "Wide Berm + 1-ft Cap" alternatives, respec- tively, which had the most and the least variation among the three sets, of any of the six Roughans Point structure combinations. 112. Assuming that for any given return period the calculated stage is a normally distributed random variable, an estimate of the probable error PE can be calculated using the three stage-frequency curves generated indepen- dently from the three sets of selected events. PE estimates the 50 percent error bounds about the mean value in a series of measurements. Equation 18 states the relationship between PE and standard deviation o as PE = 0.6745 o (18) 87 j=) > o z — ire uJ ao a = wn 25 RETURN PERIOD, YRS STAGE FREQUENCY ROUGHANS POINT INTERIOR THREE SET COMPARISON WIDE BERM Figure 58. Stage-frequency curves, three-set comparison, "Wide Berm" NTT ELPA TLE STAGE (FT NGVD) $e Saas) 5 HIN TIN 1 RETURN PERIOD, YRS STAGE FREQUENCY ROUGHANS POINT INTERIOR THREE SET COMPARISON UIDE BERM + 1FT CAP Figure 59. Stage-frequency curves, three-set comparison, "Wide Berm + 1-ft Cap" 88 Table 11 shows the relationship between the range of stages Ra calculated at any return period and oa , (Beyers 1966). Table 11 Estimate of Standard Deviation from Range Sample Size Estimate of a 2 0.8862 Ra 3 0.5908 Ra 4 0.4857 Ra 113. A single stage-frequency curve with probable error bands at se- lected return periods was produced using the following process. First, at each return period where error bounds were desired, the range of simulated stages was determined by ranking the three values and subtracting the smallest from the largest. Second, the PE was estimated using Table 11 and Equa- tion 18. Third, a single curve was produced by processing all 150 events as one set using the methods discussed in Part VI. Finally, the PE bounds were placed upon this combined curve. Figures 60 and 61 show the combined curves with error bounds which correspond to the three curves shown in Figures 58 and 59, respectively. Probable error curves are not presented for any still-water locations. The probable error of the selection process is too small to be seen for the still-water locations because of the small variability shown in Figure 57. Assessing the Impact of the Standard Project Northeaster 114. The Standard Project Northeaster (SPN) definition can be deter- mined from the definition for the Standard Project Storm (Headquarters, De- partment of the Army, Office of the Chief of Engineers, 1952) as the north- easter which results from the "most severe combinations of meteorologic" and tidal "conditions that are considered reasonably characteristic of the geo- graphical region involved, excluding extremely rare combinations." For this report two processes are important in considering the specification of an SPN, still-water level and wave overtopping. It is possible that a separate SPN would have to be defined for each process. The SPN which would produce the 89 Q > o ZS = w ul oO a = wn 10 2s RETURN PERIOD, YRS VIDE BERRA (NORTH UALL=14 FT —— —— PROBABLE ERROR LIAITS STAGE FREQUENCY ROUGHANS POINT INTERIOR PROBABLE ERROR LIMITS VIDE BERN Figure 60. Probable error of the selection process, "Wide Berm" STAGE (FT NGVD) 9 8 ? 6 5 4 3 2 i O) fo RETURN PERIOD, YRS LEGEND UIDE BERA + 1 FT (NORTH UALL*14 FT) — — —— PROBABLE ERROR LINITS STAGE FREQUENCY ROUGHANS POINT INTERIOR PROBABLE ERROR LIMITS WIDE BERM + iFT CAP Figure 61. Probable error of the selection process, "Wide Berm + 1-ft Cap" 90 highest still-water level might not produce the highest waves at Roughans Point and, therefore, not the highest overtopping rates. 115. The. SPN still-water level was estimated to be 13.0 ft NGVD (NED 1983) by adding the maximum surge recorded at Boston, about 5 ft, and the maximum probable tide, 7.4 ft NGVD and then rounding up to the next foot of elevation. This resulted in a still-water elevation which was almost 3 ft higher than the maximum ever recorded at the Boston gage. Given the unlikely event that a tide with a maximum elevation near the maximum probable tide were to occur sometime during the maximum surge producing northeaster, the proba- bility that the hour of maximum surge (using hour increments) would occur at the hour of maximum tide is only 1/24 (assuming a semidiurnal tide with un- equal highs). Consequently, this combination might fall under the "excluding extremely rare" clause in the definition of the SPN. A better specification of the SPN still-water level might be closer to 12.0 ft NGVD. 116. This report is mainly concerned with the effect of the SPN on interior flooding at Roughans Point and the propagation of the SPN still-water level throughout the study area which can be easily stated regardless of the exact specification of the SPN still-water level. In considering the interior floods at Roughans Point, the effect of an SPN is straightforward; the inter- ior of Roughans Point would fill to overflowing. The interior water level (approximately 1-2 ft higher than the still-water level in Broad Sound) would be determined by how fast the overtopping volumes would flow over roadways at the west boundary of Roughans Point. The evidence seems clear that given a water level on the order of 12-13 ft NGVD and with the waves appropriate for an SPN, all of the proposed alternatives would be swamped. This can best be seen by considering Figure 36. The only alternative which offers significant protection at the highest return periods is the "Wide Berm + 2-ft Cap." However, even this alternative would not offer protection against the SPN. The highest still-water level (in Broad Sound) tested in the simulations was 11.2 ft, roughly a 500-year level. Although the SPN would fall well to the right of the edge of Figure 36, the effect of the SPN can be estimated as follows. The extra foot of still water resulting from the SPN would change a 2-ft cap down to an effective 1-ft cap. Furthermore, the larger and longer Waves caused by the effect of deeper water in front of the structure and the higher wind speeds of the SPN would further increase the flood levels. Consequently, the interior levels caused by the SPN with the "Wide Berm + 2-ft 91 Cap" would be more severe than that shown for the "Wide Berm + 1-ft Cap" at the 500-year return period. It is possible that although the proposed im- provements at Roughans Point would offer considerable protection against lesser northeasters, the flood levels for the SPN might be higher after the improvements. Without the improvements, water will begin returning to the ocean over the north wall at approximately 11 ft NGVD. This outflow of water considerably lessens the probability of extreme interior flood levels. With the improvements, this outflow would be prevented by the increased wall heights until much higher water levels. The lack of data to ascertain the relative importance of outflow over the walls versus the outflow at the western edge of the Roughans Point area at extreme interior flood levels makes definitive conclusions difficult. 117. For the still-water locations the numerical storm surge model results showed that the Broad Sound maximum water levels produced by the ensemble storms were efficiently conveyed throughout the Saugus-Pines River system. Differences between outside and inside water levels were always small with the inside level usually slightly higher. The time-history of the SPN surge might be more peaked. This peaked profile would likely suffer more loss through the inlet and channel system, but this loss would be offset by the local wind setup of the shallower water of the flood plain (the cause of the higher interior levels during the simulations). Therefore, the predicted result of the SPN still-water level would be that the whole study area would flood to the level of the SPN in Broad Sound. Conclusions 118. Stage-frequency curves for 15 possible structure combinations at Roughans Point and for 14 still-water level locations were presented and dis- cussed. The potential error associated with each step of the procedure was discussed. A more formal determination of the probable error of the selection process was presented. Finally the estimated impact of the SPN was discussed for both interior flooding at Roughans Point and the still-water locations. 119. At Roughans Point, where flooding is caused by the overtopping of seawalls by storm waves, physical, numerical storm surge, numerical wave, flood routing, and probability models were needed. Multiple combinations of possible seawall-revetment structures were modeled. Major differences among 92 the combinations were evident at the lower return periods with the combina- tions of a wide berm revetment and a cap on the existing seawall for the east wall of Roughans Point providing the greatest protection. At higher return periods the protection differential offered by the various structure combina- tions tends to diminish. For still-water levels and wave conditions of an SPN, all structure combinations tested would be ineffective at protecting the interior of Roughans Point. Tests were run to determine a structure height for the north wall. These tests indicated that significant overtopping did not begin until the north wall structure was lowered below 13 ft. Since the existing height of the north wall is above this level at several sections, it is recommended that the revetment height be set at 13 ft with the wall height being set so that there is a transition between the existing wall heights. The only height that would be raised would be that of wall D, which would be raised to match wall C. 120. For areas where flooding is due to coastal inundation by the still-water level resulting from the combination of storm surge and astronom- ical tide, only the storm surge and probability models were necessary. These areas include both open coast and estuarine locations. For these areas flooded by the still-water level, the results of the modeling indicated that the Whole study area floods to approximately the same level. The flood levels are efficiently conveyed through the inlet and throughout the flood plain of the Saugus-Pines River system. Inside the inlet there is a small gradient in the still-water level, rising from north to south, which results from local Wind setup caused by north to northeast winds which predominate during storm conditions. This local wind setup results in flood levels inside the inlet which vary by one-half to three-fourths of a foot during the more severe storm events. Outside the river system in Broad Sound a smaller north-south gra- dient exists with differences of only a few tenths of a foot resulting. Data collected after completion of the modeling indicated that losses do occur as the flood levels are conveyed upstream of the Fox Hill Drawbridge on the Saugus River and upstream of the Highway embankment on the Pines River. Stage-frequency curves for these areas were adjusted to accommodate these additional data. The curves were lowered 0.3 and 0.5 ft at the lower return periods for upstream Saugus River and Pines river locations, respectively. These reductions were linearly reduced for higher return periods because higher flood levels would provide greater access to these areas. 93 REFERENCES Ahrens, J. P., and Heimbaugh, M. S. In preparation. "Irregular Wave Overtop- ping of Seawall/Revetment Configurations, Roughans Point Massachusetts," Technical Report, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Ahrens, J. P., and McCartney, B. L. 1975. "Wave Period Effect on the Stability of Riprap," Proceedings of Civil Engineering in the Oceans/II1, American Society of Civil Engineers, pp 1019-1034. Beyers, W. H., ed. 1966. Handbook of Tables for Probability and Statistics, The Chemical Rubber Company, Cleveland, Ohio. pp 270-272. Butler, H. L. 1978. "Coastal Flood Simulations in Stretched Coordinates," Proceedings, 16th International Conference on Coastal Engineering, Hamburg, Germany. . 1983. "Lake Pontchartrain and Vicinity Hurricane Protection Plan, Report 3, Numerical Model Investigation of Plan Impact On the Tidal Prism of Lake Pontchartrain," Technical Report HL-82-2, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Corson, W. D., et al. 1981 (Jan). "Atlantic Coast Hindcast, Deepwater, Significant Wave Information," WIS Report 2, US Army Engineer Waterways Experiment Station, Vicksburg, Miss 1982 (Mar). "Atlantic Coast Hindeast Phase II, Significant Wave Information," WIS Report 6, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Harris, D. L. 1981. "Tides and Tidal Datums in the United States," CERC Special Report No. 7, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Headquarters, Department of the Army, Office of the Chief of Engineers. 1952. "Standard Project Flood Determinations," Engineer Manual 1110-2-1411, US Government Printing Office, Washington, DC. Hubertz, J. M. 1985. "A Users Guide to a Steady-State Shallow-Water Directional Spectral Wave Model," Instruction Report CERC-86-1, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Hughes, S. A. 1984. "The TMA Shallow-Water Spectrum, Description and Appli- cations," Technical Report CERC-84-7, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Jensen, R. E. 1983. "Methodology for the Calculation of Shallow-Water Wave Climate, WIS Report 8," US Army Engineer Waterways Experiment Station, Vicks- burg, Miss. Kitaigorodskii, S. A., Krasitskii, V. P., and Zaslavakii, M. M. 1975. "On Phillips' Theory of Equilibrium Range in the Spectra of Wind Generated Gravity Waves," Journal of Physical Oceanography, Volume 5, pp. 410-420. Meyers, V. A. 1970. "Joint Probability Method of Tide Frequency Analysis Applied to Atlantic City and Long Beach Island, N. J.," ESSA Technical Memorandum WBTM HYDRO 11, US Department of Commerce, Environmental Science Services Administration, Weather Bureau, Silver Springs, Md. g4 Prater, M. D., Hardy, T. A., and Butler, H. L. In preparation. "Fire Island to Montauk Point, New York, Storm Surge Model Study," Technical Report, US Army Engineer Waterways Experiment Station, Vicksburg, Miss. Shore Protection Manual. 1984. 4th edition, 2 volumes, US Army Engineer Waterways Experiment Station, Coastal Engineering Research Center, US Govern- ment Printing Office, Washington, DC. US Army Engineer Division, New England. 1983. '"Roughans Point, Revere, Massachusetts-Coastal Flood Protection Study, Interim Report," Vols I and II, Waltham, Mass. Weggel, J. R. 1977. "Wave Overtopping Equation," Proceedings of the Fifteenth Coastal Engineering Conference, American Society of Civil Engineers, pp 2737-2755. 95 TL cee a) ‘a ' ; ; Rye Le)! ene Thr FORT | i Ml iy HOR Set Tag Mai te Re rian i ay i 7" | Lorne Nha LORIE (REL SR ike Mra F : 1 " het ny cant Ree 4c) AUER old Pray ih ey i mpeay:* Pe els et 4 wast in reer , 15) an Gh setdumatiees ry Care Diy arr ies Crab £20 Peeeahe SUEY Tht lay Cbay | beet telly ’ Si ‘dy Omid ae foata E te weve AOL eae Huey 6 BCI chee wh wz eA ore Yo one. Ps bine. , +, IG REMEL GN Tb S@iga mire tira Sings Wee et a thy Pa ti; ii oi ait t MA Sivan 5 ‘ r Reps Se ‘ { i “~~ Mi $i) Pay Med 3) te ™ ; i : ai TAA hy : 4 sia) ” i z iy } , wane 1h