Recommended Picayune Strand Restoration Plan. Roger Copp, Charles Rowney, and David Weston

Transcription

1 Use of Hydrological Modeling for Selection of the Recommended Picayune Strand Restoration Plan Roger Copp, Charles Rowney, and David Weston GR 2010

2 Presentation Outline Project Setting and Objectives Model Selection Model Development Alternatives Analysis Model Interpretation Test Model Predictions with Pilot Test Case Next Steps Lessons Learned

3 Big Cypress Basin Complex channel network with multiple flow directions depending on rainfall and gate operations High rainfall in 1995 resulted in an interbasin flow from the Imperial River to BCB Topography slopes generally to south Flows to the west are influenced by gate operations

4 What Does the xisting System Look Like? Merritt #1 Gate Drainage from Fakahatchee Strand Typical Canal Faka Union Weir #1

5 What Does the xisting System Look Like? Cypress Swamps Sawgrass Marshes Uplands with Roads to Nowhere Upland Vegetation invading Cypress Swamps

6 Objective: Re-establish Pre-Development Flow-waysways I-75 I75 Major Flowway Minor Flowway How do we recreate these flow-ways with the appropriate hydroperiod? US 41

7 First We Need a Model A standard HC-RAS surface model was tried first, but how do we build a model that allows water to flow over the land surface along pathways not pre- conceived by the modelers? Additional areas of concern included impacts to nearby partially confined well-fields, and inter-basin transfers? MIK SH/MIK 11 was chosen for the analysis

8 Model Platform - MIK SH/MIK 11 Features Integrated surface/ground water hydrology Distributed Physically based Advanced routines for hydraulic control structures Well-developed GUI xtensive application in South Florida

9 Model Development There were three versions of the MIK SH/MIK 11 model developed with varying levels of detail 1. Simple network, best available topography 2. More complex network, same topography 3. Very complex network, revised topography

10 The BODR Model used New Topography The new topography utilized elevations generated from LIDAR LIDAR does not work in heavy vegetation without extensive ground truthing The new topography worked well in areas with sparse vegetation Flow pathways in the proposed restoration area appear to be more accurate Fakahatchee Strand topography is believed to be inaccurate. LIDAR elevations are likely too high due to dense tree cover levation Difference Map LIDAR minus PIR DM Blue-green: LIDAR is lower, Wet season_1994_bodr m PIR_mean Yll Yellow-Orange: LIDAR is higher h Legend bmp_ws_ol_mea - bmp_ws_ol_mea <VALU>

11 How We Decided to Have a Complex Model 1 st model did not have US 41 culverts in the model, and modelers were not sure that the flow distribution ib i was correct 2 nd model fixed this, but the DM was determined to have problems In the 3 rd model, the number of bridges/culverts was increased from 200 to 400 Simulation time steps were cut in half to handle this Run time increased from 5 hrs/yr to 24 hrs/yr While not the best calibrated model, it was the most accurate in restoration area U.S. 41 Culverts added to model

12 Plans Started Simple, and Then Became More Complex

13 Alternatives Analysis - Iterations 1 & 2 Modeling Objective: increase water levels south of I-75 to restore wetland hydrology, decrease peak flows to tide, and maintain existing levels of flood protection upstream of project Iteration 2 pump stations were sized to maintain existing levels of flood protection north of I-75. This tripled the PS capacity Alternative Description 1. Diversion Structure Plan Divert flows just north of US Remove All Roads, Backfill Canals Two PS just south of I-75, 1-in Yr 3a. leven Plugs Plan Three PS near I-75, 2 spreader canals 3b. Spreaders, plugs, remove roads 3 PS 2 mi. S, 82 plugs, 128 mi. roads more US 41 culverts 3c. Similar to 3b Merritt PS further north, 130 mi. roads, Total PS = 860 cfs for 10-yr DS 3d. Variant of 3c (Iteration 2) 227 mi. roads, levees for private lands, Total PS = 3,800 cfs

14 What We Learned In Iterations 1 and 2 Too few blocks resulted in rapid changes in water depths at the blocks Adding more blocks resulted in a more gradual elevation change, which was closer to pre-development conditions Once-in-10-Year design storm pump stations ti were inadequate to maintain existing levels of flood protection. Impacts to private lands on the border of the restoration area were likely, requiring additional analysis

15 Alternatives Analysis Iteration 3 In this iteration, additional restoration measures were added in a variety of combinations. Does the solution have to be so structural and expensive? Alternative Description 4. Like 3D with Larger I-75 Bridges Could larger bridges mean smaller PS 6. No Pumps Plan Low tech, 46 plugs for south reaches 7. Southern Pumps Plan Pumps S of Stewart Blvd., levees 8. Merritt Plugs, Obermeyer Gates 10 Ober. Gates on Miller and Faka Union, isolate 1 mi. of FU from GW 9. Variant of 8 Ober. Gates on FU, fill Miller, Merritt, & Prairie Obermeyer Weir

16 Alternatives Analysis Iteration 3, cont d During this phase, additional restoration measures and alternative approaches were evaluated Alternative Description 10. Adjustable Weirs for All Canals No pumps/canal filling, Oberm. Gates 11. Alt 3D and Alt 9 Ober. Gates in FU, 3 PS 12. Alt 3D and Alt 9 Ober. Gates in Miller, 3 PS 13. Alt 3D and Alt 8 Ober. Gates in Miller and FU, 3 PS Various combinations of other alternatives N t t l l ti h Alt ti 6 ( ) did Non-structural solutions such as Alternative 6 (no pumps) did not work. We could keep the peak stage the same as existing, but the duration of flooding was longer with Alt. 6

17 Recommended Plan Alternative 3D MILS 0 2 I-75 Private Lands Levee 6L levee Management Roads 1250 cfs FAKA UNIO ON CANAL 2650 cfs MRRIT TT CANA AL 810 cfs PRAIRI CA NAL LMNTS 3 Spreader Canals 3 Pump Stations Canal Plugs 80% of Roads Removed Primary All Weather Roads U.S 41 POI SW levee POI N levee POI S levee Secondary Roads Flood Protection Levees

18 Analysis of Model Results for Alt 3D Overland flow depth Depth difference maps Blue is good

19 Impacts to Private Lands are Possible Simulations show slightly higher h water levels l along levees protecting private lands, however levee heights are not known Will a 0.5-foot water level rise overtop the protective levee? Overland Depth Difference, ft Levee Belle Meade farms and residential ft µ DISCLAIMR: Any information, including but not limited to software and data, received from the South Florida Water Management District ("District") in fulfillment of a request is provided "AS IS" without warranty of any kind, and the District expressly disclaims all express and implied warranties, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. The District does not warrant, guarantee, or make any representations regarding the use, or the results of the use,of the information provided to you by the District in terms of correctness, accuracy, reliability, timeliness or otherwise. The entire risk as to the results and performance of any information obtained from the District is entirely assumed by the recipient. This map is a conceptual tool utilized for project development only. This map is not self-executing or binding, and does not otherwise affect the interests of any persons including any vested rights or existing uses of real property.this is not a survey. Miles Water Depth (ft) Levee Alt3D Network

20 Modeling Results Alternative 3D Increases the Water Table South of I-75 [Ú [Ú [Ú I-75 This map shows the increase in ground water elevation for Alt 3D The 2-ft increase in ground water elevation does not cause flooding in residential areas north of I-75 Legend Mike 11 network [Ú pump station 2000groundw_wet_diffwet diff value < > 2 majrds11 Average Wet Season GW levation Difference (ft)

21 Analysis of Model Results Miller Canal WLs N of I-75 Do Not Increase with Project Alt 3D Future Without Project (cfs) Flow ( /1 8/11 8/21 8/31 9/10 9/20 9/30 15 Channel Profiles show higher WLs south of I-75 t-navd) levation (ft I-75 I75 xisting Conditions Alternative 3D FU-7 FU-6 FU-5 FU-4 FU-3 FAK FU-2 FU River Miles

22 Check Model Using Data from the Prairie Canal Test Fill The test fill of Prairie Canal began in 2005 to determine the best construction approach and to provide a real-life confirmation of the predicted hydrologic responses #* Miller Canal #* Faka Union Canal #* Merritt Ca anal Prairie Canal I #* #* #* #* U.S

23 Prairie Canal Fill - Photos

24 Measured Water Level Responses at Well Transects Miller Canal SGT1W2 SGT1W1 MLRI75 Picayune Strand_TOPO FAKI75 SGT1W4 Faka Union Canal MRTI75 LUCKLK SGT1W Measured Depth to Land Surface Transect SGT2 Post Restoration SGT2W2 SGT2W1 SGT2W3 SGT2W4 SGT2W5 feet - NGVD -2-4 SGT2W6-6 Prairie Canal -8 1/1/05 7/20/05 2/5/06 8/24/06 3/12/07 9/28/07 4/15/08 11/1/08 date SGT2W3 SGT2W6 SGT3W4SGT3W5 SGT4W1 SGT4W2 SGT4W4 SGT4W3 Merritt Canal SGT4W6 SGT3W7 SGT3W6 Legend DBHYDRO_Stage_Site SFWMD_Canals g100_topo <VALU> Well 3 (Yellow) is next to Faka Union Well 6 (Burgundy) is next to Prairie Water levels in Prairie Canal are higher after restoration

25 Simulated Results of Water Level Response to the Partial Filling of Prairie Canal Fill Simulated stages were determined for the 100- year design storm Simulated overland flow depths for the existing condition were subtracted from the filled condition The figures show the increase in overland flow depth for the proposed fill The simulated stages are higher than measured responses in Tropical Storm rnesto in Oct 2006, which is expected 2005 Fill Legend PSRP-HHP2_19_M11_HARN_ Fill Value < -0.5 ft -0.5 to -0.2 ft -0.2 to -0.1 ft -0.1 to 0.1 ft 0.1 to 0.2 ft 0.2 to 0.5 ft > 0.5 ft 2006 Fill Dark blue indicates water depth for the filled canal scenario is >0.5 ft deeper than the depth of the existing condition scenario

26 Additional Modeling to be Conducted During Design Determine if water levels during larger storms will impact private lands west and south of PSRP Flow distribution south of the pump stations Re-evaluate with better topo xisting vs Alt 3D - Bridge ft-navd /1/94 3/2/94 5/1/94 6/30/94 8/29/94 10/28/94 12/27/94 Alt 3D xisting xisting vs Alt 3D - Bridge ft-navd f /1/94 3/2/94 5/1/94 6/30/94 8/29/94 10/28/94 12/27/94 Alt 3D xisting

27 Lessons Learned From These fforts It is a good idea to start simple and only add the complexities that t you need to Data gaps and uncertainties are very difficult to handle There is a limit to how much detail you can add and how many questions you can ask of one model As you move from the concept stage to the planning and design stages, you will ask more difficult questions To answer them, you will need more advanced d tools A variety of modeling programs may be appropriate Make sure you plan adequate time to develop and calibrate the f modeling tools before you use them to make decisions Take your model for a test drive on the race track, and see if the wheels stay on