Assessment of the temporal evolution of storm surge across coastal Louisiana

Highlights

• Land to water (L:W) isopleths are derived for storm surge model mesh years 1930, 1970, 2010.

• ADCIRC storm surge model meshes are constructed featuring historic Louisiana coastal landscapes for 1930, 1970, 2010.

• Hydrologic unit code 6 (HUC6) coastal basins and HUC12 sub-watersheds are utilized to quantify and compare mesh year results.

• Results indicate riverine sediment input is a HUC6 coastal basin scale compensator for relative sea level rise.

• Reduced sediment input leads to increased storm surge heights in sediment starved coastal basins 1970-2010.

Abstract

The co-evolution of wetland loss and flood risk in the Mississippi River Delta is tested by contrasting the response of storm surge in coastal basins with varying historical riverine sediment inputs. A previously developed method to construct hydrodynamic storm surge models is employed to quantify historical changes in coastal storm surge. Simplified historical landscapes facilitate comparability while storm surge model meshes developed from historical data are incomparable due to the only recent (post-2000) extensive use of lidar for topographic mapping. Storm surge model meshes circa 1930, 1970 and 2010 are constructed via application of land to water (L:W) isopleths, lines that indicate areas of constant land to water ratio across coastal Louisiana. The ADvanced CIRCulation (ADCIRC) code, coupled with the Simulating WAves Nearshore (SWAN) wave model, is used to compute water surface elevations, time of inundation, depth-averaged currents and wave statistics from a suite of 14 hurricane wind and pressure fields for each mesh year. Maximum water surface elevation and inundation time differences correspond with coastal basins featuring historically negligible riverine sediment inputs and wetland loss as well as a coastal basin with historically substantial riverine inputs and wetland gain. The major finding of this analysis is maximum water surface elevations differences from 1970 to 2010 are 0.247 m and 0.282 m within sediment-starved Terrebonne and Barataria coastal basins, respectively. This difference is only 0.096 m across the adjacent sediment-abundant Atchafalaya-Vermilion coastal basin. Hurricane Rita inundation time results from 1970 to 2010 demonstrate an increase of approximately one day across Terrebonne and Barataria while little change occurs across Atchafalaya-Vermilion. The connection between storm surge characteristics and changes in riverine sediment inputs is also demonstrated via a sensitivity analysis which identifies changes in sediment inputs as the greatest contributor to changes in storm surge when compared with historical global mean sea level (GMSL) rise and the excavation of major navigation waterways. Results imply the magnitude of the challenge of preparing this area for future subsidence and GMSL rise.

Introduction

River engineering designs in the Mississippi River Delta have disrupted riverine sediment deposition, which normally compensates for relative sea level rise (sum of subsidence and global mean sea level (GMSL) rise) (Bentley et al., 2016; Boesch et al., 1994; Edmonds, 2012a; Louisiana Coastal Wetlands Conservation and Restoration Task Force, 1993; Paola et al., 2011; Twilley et al., 2016; Twilley et al., 2008). This disruption is negatively impacting stability of the delta landscape and threatening nationally important industries (Day et al., 2007; Syvitski et al., 2009; Vörösmarty et al., 2009). Industries at risk include a fishery that produces the largest seafood harvest in the contiguous United States, a petroleum industry that produces the most crude oil and the second most natural gas in the United States, and a shipping industry that has made Louisiana the top export state in the nation (Barnes et al., 2017; Coastal Protection and Restoration Authority, 2017; Greater New Orleans Inc., 2015; Louisiana Seafood Industry, 2017). More than 4,877 km² of coastal Louisiana wetlands along with most of the barrier islands have become submerged and converted to water from 1932 to 2010 due to, among other factors, changes in riverine sediment deposition rates and the highest measured rate of relative sea-level rise in the contiguous United States with 9.2 mm/year measured by a NOAA tide gauge in Grand Isle, LA 1947-2006 (Fig. 1a) (Batker et al., 2010; Bird, 2010; Couvillion et al., 2011; National Oceanic and Atmospheric Administration, 2018; Nienhuis et al., 2017). In addition, coastal Louisiana is susceptible to a high frequency of hurricane landfalls (Chen et al., 2008; Needham and Keim, 2012, 2014). If the current rate of wetland loss continues or accelerates with the predicted increase in the rate of GMSL rise (Jevrejeva et al., 2016; Parris et al., 2012; Stocker et al., 2013; Sweet et al., 2017), the fishing, petroleum and shipping industries will be adversely impacted, and flood risks in coastal communities will further increase. The aim of this analysis is to quantify the change in storm surge characteristics, such as surge height and inundation time, across coastal Louisiana from 1930 to 2010 and to estimate if flood risks change in coastal basins with different rates of riverine sediment inputs.

Because the impact of historical river management decisions has occurred on the coastal basin scale (Fig. 1c), hydrologic unit code 6 (HUC6) coastal basins are utilized to quantify the historical change in storm surge heights and inundation time from 1930 to 2010. Hydrologic unit codes define the entire or a partial area of a surface drainage basin with lower HUC numbers describing larger basins and higher HUC numbers describing smaller sub-watersheds. Higher HUC numbered sub-watersheds fit within the lower HUC numbered larger basins (U.S. Geological Survey, 2017a). HUC6 coastal basins (e.g. Atchafalaya-Vermilion versus Terrebonne and Barataria) are specifically treated as experimental units to quantify the rate of Gulf of Mexico (GOM) migration inland where riverine sediment input remains (sediment-abundant) and where riverine sediment input is substantially reduced (sediment-starved). HUC12 sub-watersheds, smaller watersheds that fit within the HUC6 coastal basins, are also utilized to provide a range of surge height and inundation time values within each HUC6 coastal basin. Hereafter, HUC6 coastal basins are referred to only as “coastal basins” while HUC12 sub-watersheds are referred to only as “sub-watersheds”.

Mississippi River flood control projects implemented over the past century allow the ability to contrast the evolution of storm surge in sediment-starved coastal basins versus an adjacent sediment-abundant coastal basin (Boesch et al., 1994; Twilley et al., 2008, 2016). For example, the Lafourche Delta formed approximately 1 000 to 300 years ago via Mississippi River sediment discharged into Bayou Lafourche until Bayou Lafourche was dammed at Donaldsonville in 1906 (Morgan, 1979) (Fig. 1a,c). During a 1851 flood, Mississippi River discharge into Bayou Lafourche reached 340 m³/s allowing both water and sediment to overflow the banks of Bayou Lafourche and nourish adjacent wetlands in the Terrebonne and Barataria coastal basins (Ellet, 1853). Modern Bayou Lafourche discharge is approximately 28 m³/s, which substantially limits nourishment of adjacent wetlands (Coastal Environments Inc., 1997).

In contrast, annual Mississippi River discharge (excluding Red River) diverted to the Atchafalaya increased from approximately 2.5% in 1880 to 13% in 1930 to nearly 30% in 1952 (Fisk, 1952). Suspended sediment load in the lower Mississippi River upstream of Old River pre-1963 was greater than 400 megatons (million metric tons) (Blum and Roberts, 2009; Keown et al., 1986). Red River suspended sediment load pre-1963 is estimated 35.1 MT/year and is assumed to flow directly to the Atchafalaya River. Therefore, estimated 1930 suspended sediment load at Simmesport (8 km downstream of Old River on the Atchafalaya) is 87.1 MT/year (Fig. 1c). The U.S. Army Corps of Engineers completed the Old River Control Structure in 1963 which controlled Mississippi and Red River discharge to the Atchafalaya at 30% resulting in the formation of Wax Lake and Atchafalaya Deltas (Edmonds, 2012b; Reuss, 2004; Wellner et al., 2005; Wells et al., 1984). Dams and other river improvements were also implemented along Mississippi River tributaries post-1963 which caused a gradual decline in Mississippi suspended sediment loads (Bentley et al., 2016). Keown et al. (1986) established the suspended sediment load 1970-1978 at Simmesport 93.9 MT/year. Blum and Roberts (2009) measured suspended sediment loads at Simmesport of 86 MT/year (1977) and 49 MT/year (2006) while Mize et al. (2018) measured suspended sediment loads at Melville (48 km south of Old River) of 39 MT/year (2005) and 33 MT/year (2015). The 2015 recording at Melville continues to be substantially greater than the negligible Mississippi River suspended sediment load entering Bayou Lafourche.

In the present analysis, a series of hydrodynamic storm surge model meshes are developed and feature comparable representations of historical Louisiana coastal landscapes to examine the effects of negligible riverine sediment deposition in the Terrebonne and Barataria coastal basins and substantial sediment deposition in the Atchafalaya-Vermilion coastal basin (Fig. 1a,c). Constructing storm surge model meshes featuring historical coastal landscapes circa 1930-2010 with comparable topo-bathymetric detail is not possible as contemporary data collection methods, such as lidar, has been used extensively for topographic mapping since only the early 2000s. Therefore, the method developed by Siverd et al. (2018) is employed to simplify the modern (2010) Louisiana coastal landscape, as well as to construct comparable, historical landscapes. This method entails application of land to water (L:W) isopleths which are defined as lines on a map connecting points of constant value of the land to water ratio (Gagliano et al., 1970; Siverd et al., 2018; Twilley et al., 2016). A simplified coastal landscape featuring a L:W isopleth permutation of 99%-90%-40%–1% with coastal zones labeled high (99%–90%), intermediate (90%–40%) and submersed (40%–1%) was found to most closely reproduce the detailed coastal Louisiana landscape (Siverd et al., 2018).

The goal of the present analysis is to develop storm surge model meshes featuring historical landscapes (i.e. 1930, 1970 and 2010) via the same method established by Siverd et al. (2018) and to quantify changes in storm surge characteristics over this 80-year period of change in riverine sediment deposition. A secondary goal is to conduct sensitivity analyses to find the largest contributor to increased inland surge heights (L:W isopleth migration, addition of deep navigation waterways or GMSL rise). Results include the quantification of changes in storm surge characteristics (e.g. maximum of maximums water surface elevations and inundation time) within coastal basins and smaller sub-watersheds. These two surge characteristics are then associated with change in historical riverine sediment input rates and subsequent wetland loss/gain across coastal basins.