remote sensing Article
InSAR-Based Mapping of Tidal Inundation Extent and Amplitude in Louisiana Coastal Wetlands Talib Oliver-Cabrera * and Shimon Wdowinski Division of Marine Geosciences, University of Miami, Miami, FL 33149, USA;
[email protected] * Correspondence:
[email protected]; Tel.: +1-305-421-4928 Academic Editors: Javier Bustamante, Alfredo R. Huete and Prasad S. Thenkabail Received: 29 February 2016; Accepted: 3 May 2016; Published: 7 May 2016
Abstract: The Louisiana coast is among the most productive coastal areas in the US and home to the largest coastal wetland area in the nation. However, Louisiana coastal wetlands have been disappearing at an alarming rate due to natural and anthropogenic processes, including sea level rise, land subsidence and infrastructure development. Wetland loss occurs mainly along the tidal zone, which varies in width and morphology along the Louisiana shoreline. In this study, we use Interferometric Synthetic Aperture Radar (InSAR) observations to detect the extent of the tidal inundation zone and evaluate the interaction between tidal currents and coastal wetlands. Our data consist of ALOS and Radarsat-1 observations acquired between 2006–2011 and 2003–2008, respectively. Interferometric processing of the data provides detailed maps of water level changes in the tidal zone, which are validated using sea level data from a tide gauge station. Our results indicate vertical tidal changes up to 30 cm and horizontal tidal flow limited to 5–15 km from open waters. The results also show that the tidal inundation is disrupted by various man-made structures, such as canals and roads, which change the natural tidal flow interaction with the coast. Keywords: InSAR; wetlands; tidal inundation; remote sensing
1. Introduction Louisiana has more coastal wetlands than any other state in the US, covering an area of roughly 20,000 km2 . These coastal wetlands provide important ecosystem services including habitat for a wide variety of wildlife, storm protection, flood control, water storage and others. Louisiana coastal wetlands also have high economical values, including habitat for fishing industry, recreational activities and location for oil and gas exploitation. The excessive use of the ecosystem services, climatic drivers, and human sprawl have threaten these valuable wetlands [1]. Coastal Louisiana has undergone a decrease in land area of approximately 4747 km2 in the period between 1932 and 2010, reaching a rate of 42.9 km2 of yearly wetland loss between 1985 to 2010 [2]. To protect these prolific resources, it is important to monitor and quantify the circumstances that provoke wetland loss in coastal Louisiana. Monitoring coastal wetlands is a difficult task due to their large extent and hostile environment, which limit ground-based efforts. Remote sensing observations provide a suitable solution for analyzing large areas in a safe and economic manner. Several remote sensing studies were conducted to analyze and monitor Louisiana’s coastal wetlands. Penland et al. [3] used aerial photography to describe land loss of the Mississippi River delta plain and explained the loss by three main processes: erosion, submergence and direct removal. Steyer et al. [4] monitored vegetation changes by generating vegetation indices from MODIS and Landsat imagery, using Normalized Difference Vegetation Index (NDVI) to quantify the impacts of hurricanes Katrina and Rita (2005) on the coastal marshes of Louisiana. Barras et al. [5] used aerial photography and Landsat images to estimate land loss of coastal Louisiana; they also projected the land loss to 2050. Ramsey et al. [6] analyzed L-band and C-band Synthetic Aperture Radar (SAR) amplitude combined with Landsat imagery to study flooding in Remote Sens. 2016, 8, 393; doi:10.3390/rs8050393
www.mdpi.com/journal/remotesensing
Remote Sens. 2016, 8, 393
2 of 13
coastal marshes of Louisiana. They created marsh inundation maps by comparing SAR backscatter from marshes during flooding and non-flooding conditions. A detailed description of the use of the different remote sensing techniques for wetland observations was recently summarized by Gallant [1]. The above studies provide a good assessment of vegetation or land changes. However, they have limited capability to detect water level changes in tidal zone, where active land loss occurs. Coastal wetlands and water levels are interrelated, as water level changes may induce shifts that produce disturbances or environmental changes [7]. A different remote sensing method that can detect water level changes in wetlands is Interferometric Synthetic Aperture Radar (InSAR). Studies have shown that InSAR observations can detect water level change with 3–5 cm accuracy level over wide wetland areas (e.g., [8–11]). InSAR measures phase difference between two SAR acquisitions from roughly the same location in space. The result generated from the phase-difference operation is a phase change map, called an interferogram, which is sensitive to changes of the Earth’s surface [12]. Interferograms are widely used to measure surface displacements after an earthquake, urban subsidence due to ground water withdrawal, magmatic-induce deformation and, as in the case of this research, water-level changes in wetlands. Most wetland InSAR studies were applied to inland freshwater wetlands, where hydrological changes occur within days or weeks. Several studies showed that the technique works well in woody wetlands [11,13]. Only few studies applied the technique to study water level changes in coastal areas. Ramsey et al. [14] analyzed four tandem ERS-1/2 interferograms and noticed phase changes between freshwater and saltwater vegetation. Wdowinski et al. [8] also noticed phase change occurring along the transition between fresh and salt-water vegetation in the western Everglades (Florida). Recently, Wdowinski et al. [15] observed tide inundation in the coastal wetlands of the Florida Everglades, showing the feasibility of InSAR for detecting tidal inundation propagation and mapping the tidal zone. In this study we use Synthetic Aperture Radar (SAR) data acquired by the ALOS and Radarsat-1 satellites as well as tide gauge data to map the tidal inundation extent and water level changes throughout the coastal wetlands of Louisiana. We were able to detect water level changes up to 30 cm in some areas. Our observations also show disruptions of water inlet into the wetlands due to man-made structures. 2. Louisiana Coastal Wetlands Louisiana’s coastal wetlands are located in the southern part of the state, along the entire state shoreline. Sedimentation processes induced by the Mississippi river dominate the morphology of the coast. The shift of the river’s deposition plume formed two main features, the Chenier Plain and the Mississippi Deltaic Plain [16] (Figure 1). The Chenier Plain is located along the western part of the state, where the coastline’s morphology consists of wooded beach ridges termed Cheniers and mudflats covered by swamps and marshes [17]. The Mississippi Deltaic Plain is located along the eastern shoreline, where the coastal morphology is a product of sediment deposition. It is also the present location of the Mississippi river delta. Wetland vegetation varies along the coast depending mostly on water conditions due to mixed inputs of fresh and salt-water within the coastal zone. Two main types of wetland vegetation are found along the Louisiana coast, forested wetlands, termed swamps, and herbaceous wetlands, termed marshes [18]. Swamps are mostly found in freshwater zones whereas most marshes are located in brackish or saline water. The map in Figure 1 shows a vegetation transition along Louisiana’s coast. The vegetation type works roughly as a marker of the extents of salt-water inputs, for example, the presence of fresh-water marshes in areas that are not affected by ocean tide. The areas on the map with color green represent vegetation that is better adapted to the salt-water intrusions. Meanwhile the purple and orange represent vegetation that requires fresh-water inputs to live. The transition zone between fresh and
Remote Sens. 2016, 8, 393
3 of 13
salt water is marked by the presence of brackish marsh (aqua green color) and intermediate marsh (purple color). Remote Sens. 2016, 8, 393
3 of 13
Figure 1. Location Location map and wetland classification generated using Gulf-Wide Information System (G-WIS) [19], Landsat mosaic mosaic [20] as background background and and Natural Natural Earth Earth data data used used for for the the location location map. map. The wetlands wetlandswith withdotdot pattern correspond the Chenier Plain hatch to the pattern correspond to thetoChenier Plain and the and hatchthe pattern to pattern the Mississippi River Delta River plain. Delta plain. Mississippi
Louisiana and Louisiana coastal coastal wetlands wetlands have have been been threatened threatened by by both both natural natural (e.g., (e.g., sea sea level level rise) rise) and 2 anthropogenic (e.g., infrastructure development) processes causing an average loss of 43 km anthropogenic (e.g., infrastructure development) processes causing an average loss of 43 km2 of of wetlands per year [2]. [2]. Boesch the change change from from vegetated vegetated wetlands per year Boesch et et al. al. [16] [16] define define the the loss loss of of wetlands wetlands as as the wetlands to to uplands or drained wetlands uplands or drained areas areas and and to to nonvegetated nonvegetated wetlands wetlands or or submerged submerged habitats. habitats. Between Between 22 of wetlands that were transformed to open water [21]. 1956 and 2004 the state lost around 3000 km 1956 and 2004 the state lost around 3000 km of wetlands that were transformed to open water [21]. According to to Penland Penland et et al. al. [3] [3] between 1932 and and 1990, According between 1932 1990, 26% 26% of of the the coastal coastal land land loss loss was was due due to to natural natural erosion processes, 36% due to oil and gas extraction, and 23% was due to altered hydrology product erosion processes, 36% due to oil and gas extraction, and 23% was due to altered hydrology product of channel construction construction for for navigation navigation or or levees. levees. Furthermore construction of multiple multiple activities activities as as channel Furthermore the the construction of flood control levees has interrupted the normal sediment deposition from the Mississippi of flood control levees has interrupted the normal sediment deposition from the Mississippi River, River, which created created aa deficit deficit of of deposited deposited material, material, and and resulted resulted in in land land loss loss [22]. [22]. The which The vegetation vegetation in in the the submerged submerged areas areas dies dies or or has has aa limited limited growth growth due due to to hypoxia, hypoxia, transforming transforming the the vegetated vegetated areas areas into into tidal [23]. tidal mudflats mudflats [23]. 3. 3. Data Dataand andData DataProcessing Processing This This study study is is based based on on data data acquired acquired by by two two SAR SAR satellites, satellites, Advanced Advanced Land Land Observing Observing Satellite Satellite (ALOS) and Radarsat-1 (Table S1). ALOS was launched by the Japanese Space Agency (JAXA) (ALOS) and Radarsat-1 (Table S1). ALOS was launched by the Japanese Space Agency (JAXA) and and operated had a operated from from 2006 2006 to to 2011. 2011. ItItwas wasequipped equippedwith withL-band L-band(23.62 (23.62cm cmwavelength) wavelength)radar radarand and had repeated orbit of of 4646 days. Radarsat-1 waswas launched by the Canadian Space Agency in 1995. It was a repeated orbit days. Radarsat-1 launched by the Canadian Space Agency in 1995. It equipped with C-band (5.66 cm wavelength) radar and hadand a repeat resolution of 24 daysof(Table 1). was equipped with C-band (5.66 cm wavelength) radar had aorbit repeat orbit resolution 24 days We used acquired both satellites the entire of coastal (Table 1). all Weavailable used all data available databy acquired by bothover satellites over length the entire length Louisiana of coastal (Figure 2) that are archived in the User Remote Sensing Access (URSA) at the Alaska Satellite Facility Louisiana (Figure 2) that are archived in the User Remote Sensing Access (URSA) at the Alaska Satellite (ASF) data pool [24]. We used a total of 246 ALOS scenes and 154 Radarsat-1 scenes (Table 1). Facility (ASF) data pool [24]. We used a total of 246 ALOS scenes and 154 Radarsat-1 scenes (Table 1).
RemoteSens. Sens.2016, 2016,8,8, 393 Remote
Satellite ALOS Satellite ALOSALOS ALOSALOS ALOSALOS ALOSALOS ALOS ALOSALOS ALOSALOS Radarsat-1 Radarsat-1 Radarsat-1 Radarsat-1
of13 13 44of
Table 1. Scenes available and satellite information. Table 1. Scenes available and satellite information. Path Frames Orbit Number of Scenes Date 164Path 570 and 580 Ascending 36 2006–2011 Frames Orbit Number of Scenes Date 165164 570 and 580 Ascending 40 2007–2011 570 and 580 Ascending 36 2006–2011 166165 570 and 580 Ascending 40 2007–2010 570 and 580 Ascending 40 2007–2011 570 and Ascending 40 38 2007–2010 167166 570 and 580580 Ascending 2007–2011 Ascending 38 50 2007–2011 570 and 168167 570 and 580580 Ascending 2007–2011 168 570 and 580 Ascending 50 2007–2011 169169 580 580 Ascending 2006–2011 Ascending 18 18 2006–2011 170170 580 580 Ascending 2006–2011 Ascending 24 24 2006–2011 74 and Ascending 72 72 2002–2007 94 94 74 and 75 75 Ascending 2002–2007 74 and 75 Ascending 82 2002–2008 337337 74 and 75 Ascending 82 2002–2008
Figure 2. Location map of the Synthetic Aperture Radar (SAR) scenes and the available National Figure 2. Location map ofAdministration the Synthetic (NOAA) Aperturetide Radar (SAR) scenes the available National Oceanic and Atmospheric gauge stations forand Louisiana (triangles). The Oceanic and Atmospheric Administration (NOAA) tide gauge stations for Louisiana (triangles). The orange triangle marks the location of the tide gauge used in this study. orange triangle marks the location of the tide gauge used in this study.
InSAR data processing aims at detecting phase difference between two acquisitions to generate InSAR data processing aims at detecting phase difference between two acquisitions to generate interferograms, reflecting surface changes between the two acquisitions. InSAR is regularly used to interferograms, reflecting surface changes between the two acquisitions. InSAR is regularly used to measure displacements of solid surfaces. However, in the case of wetlands, coherent interferometric measure displacements of solid surfaces. However, in the case of wetlands, coherent interferometric signal occurs due to double bounce scattering, from the water surface and vegetation, to measure signal occurs due to double bounce scattering, from the water surface and vegetation, to measure water level changes [25,26]. Data was processed using the ROI_PAC package [27], with 4–8 looks in water level changes [25,26]. Data was processed using the ROI_PAC package [27], with 4–8 looks in order to co-register pairs of images. Co-registration searches similarities in morphology between two order to co-register pairs of images. Co-registration searches similarities in morphology between two acquisitions and the looks are an average of pixels with the similar coherence. A lower threshold acquisitions and the looks are an average of pixels with the similar coherence. A lower threshold of of looks provides higher resolution results. However, due to the lack of topography in the area, the looks provides higher resolution results. However, due to the lack of topography in the area, the use use of a lower threshold of looks does not permit proper co-registering of the SAR images. Unlike of a lower threshold of looks does not permit proper co-registering of the SAR images. Unlike most most InSAR applications, we evaluated both filtered and unfiltered results and we didn’t unwrap the InSAR applications, we evaluated both filtered and unfiltered results and we didn’t unwrap the interferograms. As wetland InSAR phase changes tend to be localized, we avoided these commonly interferograms. As wetland InSAR phase changes tend to be localized, we avoided these commonly used used processing steps, which can affect localized signal in order to increase signal to noise ratio. processing steps, which can affect localized signal in order to increase signal to noise ratio. Fringes represent phase changed due to water level changes in the interferograms, in which Fringes represent phase changed due to water level changes in the interferograms, in which each each fringe corresponds to a distance change of λ/2, where λ is the wavelength of the SAR sensor. fringe corresponds to a distance change of λ/2, where λ is the wavelength of the SAR sensor. Because Because phase changes in wetlands predominantly reflect vertical movement of the surface, the line of phase changes in wetlands predominantly reflect vertical movement of the surface, the line of sight sight (LOS) measurement can be translated to vertical movement, by accounting for the acquisition (LOS) measurement can be translated to vertical movement, by accounting for the acquisition geometry. The translation from LOS to vertical movement is inversely proportional to the cosine of the geometry. The translation from LOS to vertical movement is inversely proportional to the cosine of acquisition’s incidence angle [28]. Based on interferogram interpretation, a vector line representing the the acquisition’s incidence angle [28]. Based on interferogram interpretation, a vector line representing the tidal inundation extents along the coast of Louisiana was digitized. The InSAR detected extents follow the observed phase changes and disruptions along the coast.
Remote Sens. 2016, 8, 393
5 of 13
Remote Sens. 2016, 8, 393 tidal inundation extents
5 of 13 along the coast of Louisiana was digitized. The InSAR detected extents follow the observed phase changes and disruptions along the coast. Previous Previous studies studies indicate indicate that that the the coherence coherence of of wetland wetland interferograms interferograms is is very very sensitive sensitive to to the the temporal and perpendicular baselines (e.g., [11,13]). However, the temporal baseline contribution temporal and perpendicular baselines (e.g., [11,13]). However, the temporal baseline contribution varies varies depending depending on on the the sensor sensor and and wetland wetland types. types. Kim Kim et et al. al. [11] [11] showed showed that that fresh fresh water water wetlands wetlands with woody vegetation keep better coherence than herbaceous wetlands, mainly because tree with woody vegetation keep better coherence than herbaceous wetlands, mainly because tree trunks trunks tend tend to to be be stable stable scatterers scatterers over over time, time, whereas whereas soft-stemmed soft-stemmed vegetation vegetation are are altered altered by by vegetation vegetation growth growth and and wind. wind. In Inorder orderto toobtain obtainhigh highcoherence coherenceinterferograms, interferograms, we wegenerate generateinterferograms interferograms with with short temporal baselines, 46–92 days for ALOS data and 24–48 days for Radarsat-1 short temporal baselines, 46–92 days for ALOS data and 24–48 days for Radarsat-1 data. data. In In addition addition to to the the satellite satellite data, data, we we also also used used tide tide gauge gauge records records from from Port Port Fourchon Fourchon station station (Figure 2) installed and operated by the National Oceanic and Atmospheric Administration (NOAA). (Figure 2) installed and operated by the National Oceanic and Atmospheric Administration (NOAA). The The tide tide gauge gauge data data is is sampled sampled every every 66 min min and and is is available available online online [29]. [29].
4. 4. Results Results We ALOS andand twotwo Radarsat-1 tracks and generated an average of 10 We analyzed analyzedaatotal totalofofseven seven ALOS Radarsat-1 tracks and generated an average interferograms per track temporal baseline ranging between 24–48 days for Radarsat-1 and 46–92 for of 10 interferograms perwith track with temporal baseline ranging between 24–48 days for Radarsat-1 ALOS. We present here eight representative interferograms, some were combined into a mosaic mapa and 46–92 for ALOS. We present here eight representative interferograms, some were combined into that covers the extent of Louisiana’s coastline (Figure 3). The rest of the interferograms are presented mosaic map that covers the extent of Louisiana’s coastline (Figure 3). The rest of the interferograms in supplementary materials (Figure S1). The mosaic patterns overpatterns both inland arethe presented in the supplementary materials (Figure S1).map Theshows mosaicfringe map shows fringe over and coastal areas. The inland fringe patterns are mostly a result of water level changes in wetlands both inland and coastal areas. The inland fringe patterns are mostly a result of water level changes or likeactivities, the oneslike on the tracksarea 168ofand 167.168 Long in agriculture wetlands oractivities, agriculture the central ones onarea the of central tracks andwavelength 167. Long fringes (20–40 km) in(20–40 the north ofnorth the interferogram from track 165 andtrack 168 are likely wavelength fringes km) part in the part of the interferogram from 165most and 168 arephase most changes due to atmospheric effects [10]. Fringe discontinuities occur along swath edges, because the likely phase changes due to atmospheric effects [10]. Fringe discontinuities occur along swath edges, selected are from different time periods represent distinct signal response because interferograms the selected interferograms are from differentand, timehence, periods and, hence, represent distinct according to phase changes of each interferogram. We processed ALOS data of the entire coastline. signal response according to phase changes of each interferogram. We processed ALOS data of the However, the produced were unable towere detect details from the Chenier coast. entire coastline. However,interferograms the produced interferograms unable to detect details from plain the Chenier Thus present has a shorter over the Chenier plain plain we coast. ThusRadarsat-1, we presentwhich Radarsat-1, which wavelength has a shorterinterferograms, wavelength interferograms, over the and ALOS interferograms over the Mississippi River delta. Chenier plain and ALOS interferograms over the Mississippi River delta.
Figure 3. 3. Mosaic Mosaicof ofRadarsat-1 Radarsat-1(left (leftside) side)and andALOS ALOS(right (rightside) side)interferograms interferogramsfor forcoastal coastalLouisiana. Louisiana. Figure The red frames show the location of enlarged interferograms presented in Figures 4 and 5. The red frames
The L-band ALOS interferograms show better coherence than those obtained using C-band Radarsat-1. This is a result of the interaction between the different wavelength of each satellite and the vegetation [11]. Radarsat-1 produces less coherent results because its shorter wavelength (C-band 5.6 cm) interacts with leaves and soft stem vegetation, which change rapidly with time. The L-band
Remote Sens. 2016, 8, 393
6 of 13
The L-band ALOS interferograms show better coherence than those obtained using C-band Radarsat-1. This is a result of the interaction between the different wavelength of each satellite and theRemote vegetation [11]. Radarsat-1 produces less coherent results because its shorter wavelength (C-band Sens. 2016, 8, 393 6 of 13 5.6 cm) interacts with leaves and soft stem vegetation, which change rapidly with time. The L-band ALOS (24(24 cm) signal and interacts interactswith withbranches branches and tree trunks, ALOS cm) signalpenetrates penetratesthrough through tree tree canopies canopies and and tree trunks, which have fewer changesover overtime. time. which have fewer changes Coastalwetlands wetlandsphase phasechanges changes are are characterized characterized by require Coastal by short shortwavelength wavelengthpatterns patternsand and require zooming into some specific areas (Figures 4 and 5). Interferograms from the western most part of of zooming into some specific areas (Figures 4 and 5). Interferograms from the western most part coastal Louisiana, the Chenier Plane, show a group of fringe discontinuities aligned with the coastline coastal Louisiana, the Chenier Plane, show a group of fringe discontinuities aligned with the coastline and relate to the coastal geomorphology in that (Figure 4a,d). Chenier Plane is a zone of and relate to the coastal geomorphology in that areaarea (Figure 4a,d). TheThe Chenier Plane is a zone of ridges ridges aligned parallel to the coast; the ridges are 20 to 30 km long and 2 to 6 m high with large aligned parallel to the coast; the ridges are 20 to 30 km long and 2 to 6 m high with large mudflats mudflats covered with wetland vegetation [30]. InSARalso results also show due to disruptions of covered with wetland vegetation [30]. InSAR results show that duethat to disruptions of the tidal the tidal flow along the west Chenier Plain, the extent of tidal inundation ranges between 800 m to 2 flow along the west Chenier Plain, the extent of tidal inundation ranges between 800 m to 2 km. In the km. In the center and east of the Chenier Plain, both ALOS and Radarsat interferograms show center and east of the Chenier Plain, both ALOS and Radarsat interferograms show noticeable phase noticeable phase discontinuities on different locations along the coastline. The mapped tidal zone in discontinuities on different locations along the coastline. The mapped tidal zone in this area extends this area extends between 2–4 km (Figures 4a,b and 5d). For the Mississippi River delta, ALOS results between 2–4 km (Figures 4a,b and 5d). For the Mississippi River delta, ALOS results show fringe show fringe changes due to both natural tidal inundation and anthropogenic activity (Figure 5). changes due to both natural tidal inundation and anthropogenic activity (Figure 5). Fringe changes Fringe changes due to tidal inundation are most apparent west of the Mississippi delta, where the due to tidal inundation apparent delta, shows interferogram shows are twomost fringe cycles, west 5 kmoftothe theMississippi west side and 15where km tothe theinterferogram east side, of tidal two fringe cycles, 5 km to the the coast west (Figure side and 15The km observed to the eastdiscontinuities side, of tidal inundation inundation change over 5g). for this area change coincideover withthe coast (Figure 5g). The observed discontinuities for this area coincide with man-made structures that man-made structures that where present along the Mississippi River delta (Figure 5a,b). Our results where present River delta 5a,b). Our results thatshows from the total indicate thatalong from the the Mississippi total 743 km length of (Figure the Louisiana coast, tidal indicate inundation to be 743disrupted km length the Louisiana coast, tidal inundation to be disrupted by man-made canals byofman-made canals in roughly 45% of the shows coast and levees disruptions are observed in in roughly of thelength. coast and levees disruptions are observed in 15% of the coast length. 15% of45% the coast
Figure 4. 4.(a) thecentral central part of the Chenier spanning 24 (18 days Figure (a)Radarsat-1 Radarsat-1 interferograms interferograms ofofthe part of the Chenier PlainPlain spanning 24 days (18January January2007–11 2007–11 February 2007). The white square shows the location of figures (b,c); (b) Zoom February 2007). The white square shows the location of figures (b,c); (b) Zoom in in showing showing fringe fringe discontinuities discontinuitiessub-parallel sub-paralleltotothe thecoastline coastlinethat thatcoincides coincideswith withman-made man-made canals canals observed ininthe earth image imageofofthe thesame samearea; area;(d)(d)Radarsat-1 Radarsat-1 observed theGoogle GoogleEarth Earthimage image (c); (c); (c) (c) Google Google earth interferogram thewestern westernChenier ChenierPlain Plain spanning spanning 24-day 2004–27 January 2004). interferogram ofof the 24-dayperiod period(3(3January January 2004–27 January 2004). central of the interferogram fringe patterns show changes in inland water bodies. In In thethe central partpart of the interferogram fringe patterns show changes in inland water bodies. TheThe white whiteonsquare onside the left side of the interferogram showspatterns fringe patterns to the morphology the square the left of the interferogram shows fringe due to due the morphology of the of Chenier Chenier reflecting water level changes in the mudflats in between the ridges of the Chenier’s; (e) reflecting water level changes in the mudflats in between the ridges of the Chenier’s; (e) Google earth Google earth image of the same study area. image of the same study area.
Remote Sens. 2016, 8, 393 Remote Sens. 2016, 8, 393
7 of 13 7 of 13
Figure 5. (a) ALOS interferograms of the Mississippi Delta showing phase change due to water level Figure 5. (a) ALOS interferograms of the Mississippi Delta showing phase change due to water level changes, spanning 46 46 days between acquisitions (16(16 November 2010–1 January 2011); (b)(b) Zoom in in area changes, spanning days between acquisitions November 2010–1 January 2011); Zoom showing phase change due to tide inundation disruption caused by the presence of levees along area showing phase change due to tide inundation disruption caused by the presence of levees alongthe river; (c) Google earthearth image of the same study area; ofthe thecentral central Chenier the river; (c) Google image of the same study area;(d) (d)ALOS ALOS interferogram interferogram of Chenier Plain showing phase changes due to water level changes, spanning 92 days (3 July 2007–3 October Plain showing phase changes due to water level changes, spanning 92 days (3 July 2007–3 October 2007); 2007); (e) Elongated fringe patterns sub-parallel the coastline coinciding man-made (e) Elongated fringe patterns sub-parallel to thetocoastline coinciding with with man-made canals;canals; (f) (f) Google GoogleEarth Earthimage imageofofthe thesame samestudy studyarea; area;(g) (g)ALOS ALOSinterferogram interferogram the west the Mississippi ofof the west ofof the Mississippi River delta, spanning 92-day period (11 September 2007–12 December 2007). The interferogram shows River delta, spanning 92-day period (11 September 2007–12 December 2007). The interferogram fringe patterns to tidal level changes; Google image of image the same study area. shows fringedue patterns dueinduced to tidalwater induced water level (h) changes; (h)earth Google earth of the same study area.
5. Discussion 5. Discussion The use of L-band and C-band sensors allowed us to detect different type of tide inundation The use of L-band C-banddue sensors allowed us toin detect differentand typerepeat of tidepass inundation disruptions along coastaland Louisiana to the difference wavelength time. We disruptions along coastal Louisiana due to the difference in wavelength and repeat pass time. We focus our analysis on short wavelength changes, because the interaction between the tide inundation focus our analysis on short wavelength changes, because the interaction between the tide inundation and land or man-made structures happens in localized areas over a few kilometer distance. The and land or man-made structures happens in localized areas over a few kilometer distance. The Chenier Plane is characterized by phase changes caused by three main causes, which are ridge Chenier Plane is characterized by phase changes caused by three main causes, which are ridge morphology, inland water bodies, and man-made structures (Figure 4). For this area we found that morphology, inland water bodies, and man-made structures (Figure 4). For this area we found that both both ALOS (L-band) and Radarsat-1 (C-band) results show phase disruptions that correlate with ALOS (L-band) and Radarsat-1 (C-band) results show phase disruptions that correlate with numerous numerous man-made canals, mainly dredged for oil or recovery or navigation purposes man-made canals, dredged formainly oil recovery navigation purposes (Figures 4b(Figures and 5e).4b and 5e). However, Radarsat-1 alsovery show verywavelength short wavelength fringe patterns to the However, Radarsat-1 results results also show short fringe patterns parallel parallel to the coast coast due to the geomorphology of the Cheniers (Figure 4d). The geomorphology of the Chenier due to the geomorphology of the Cheniers (Figure 4d). The geomorphology of the Chenier has beenhas been affected anthropogenic activities including mining, constructions and colonization affected by by anthropogenic activities including sandsand mining, roadroad constructions and colonization as as described describedin in the the Cheniers Cheniersand andNatural NaturalRidges RidgesStudy Studyofofthe theState StateofofLouisiana Louisiana(2009). (2009). The short The short wavelength of of Radarsat-1 (5.66 cm) enables wavelength Radarsat-1 (5.66 cm) enablesthe thesatellite satellitetotodetect detectphase phasechanges, changes, where where ALOS ALOS failed to detect. The subtle phase changes induced by the Chenier geomorphology are too to detect. The subtle phase changes induced by the Chenier geomorphology are toosmall smallfor forthe the24 24cm cm wavelength of L-band to be detected. Two main types changes, of phase tide changes, tide inundation wavelength of L-band to be detected. Two main types of phase inundation disruptions disruptions waterthrough level changes thecharacterize tidal flush the zone characterize the area of the and water leveland changes the tidalthrough flush zone area of the Mississippi River delta. ALOS interferograms were able to detect short wavelength tidal flow disruptions, similar to the ones
Remote RemoteSens. Sens.2016, 2016,8,8, 393
8 8ofof1313
Mississippi River delta. ALOS interferograms were able to detect short wavelength tidal flow indisruptions, Radarsat-1 similar results, mainly due to construction flood protection (Figure 5a). Interferograms to the ones inlevee Radarsat-1 results,formainly due to levee construction for flood of the western side of the Mississippi River delta show a group of elongated fringes that amap the protection (Figure 5a). Interferograms of the western side of the Mississippi River delta show group tidal flush zone in that region 6 and Sea level the6tidal zone are of elongated fringes that map (Figures the tidal 5g, flush zone7a). in that regionchanges (Figuresin5g, and flush 7a). Sea level detected because it is an emerged vegetation environment with changing water levels. Although changes in the tidal flush zone are detected because it is an emerged vegetation environment with tide-induced water levelAlthough changes occur also further the open water, lack of vegetation changing water levels. tide-induced watersouth level in changes occur alsothe further south in the prevents a coherent interferometric in such environment (Figure 6).signal Further the tidal flush open water, the lack of vegetationsignal prevents a coherent interferometric in north such environment zone is delimited bynorth a man-made interrupts the tidal the path to its (Figure 6). Further the tidalcanal flushthat zone is delimited by ainundation, man-made obstructing canal that interrupts the natural limit in higher lands. the path to its natural limit in higher lands. tidal inundation, obstructing
Figure6.6.Schematic Schematicdiagram diagramshowing showing topography, vegetation induced change Figure thethe topography, vegetation andand tidaltidal induced levellevel change along along transect A-A’. transect A-A’.
The highest fringe gradient was found along the coastal wetlands in Timbalier bay, Caminada fringe was found along thedelta coastal wetlands Timbalier bay, Caminada bay The and highest Barataria bay, gradient west of the Mississippi River (Figures 5g,hinand 7a). However, not all bay and Barataria bay, west of the Mississippi River delta (Figures 5g,h and 7a). However, not interferograms of the same area show the same amount of fringe changes. For example, the 46-day all interferograms of the same area show the same amount of fringe changes. For example, the interferogram of 16 March 2009–1 May 2009 shows no phase change between the acquisitions (Figure 7b). 46-day interferogram of 16 March 2009–1 May 2009 shows no phase change theaacquisitions In contrast the 11 September 2007–12 December 2007 interferogram (Figurebetween 7a) shows set of two (Figure 7b). In contrast the 11 September 2007–12 December 2007 interferogram (Figure 7a) shows fringe cycles, which represent a change of 30 cm of vertical motion occurring between the open watera set fringe cycles, whichfurther represent a change of 30 of verticalthis motion occurring between the of of thetwo Gulf and the wetlands inland. In order to cm understand inconsistent fringe pattern, open water of the Gulf and the wetlands further inland. In order to understand this inconsistent fringe we compared the InSAR results with the tide gauge information located closest to the area of interest. pattern, we compared InSAR results with located to thecan areabeof Due to the tide gaugethe station distribution andthe thetide fewgauge placesinformation where natural tidalclosest inundation interest. Due the tide gauge station distribution the results few places natural inundation observed, thetocomparison between tide gauges andand InSAR was where conducted onlytidal for the results can be observed, the comparison between tide gauges and InSAR results was conducted only for the of ALOS track 166 and data acquired by the Port Fourchon station (Figure 7a,b). We use a 24-h period results ALOS track 166each andacquisition, data acquired byGreenwich the Port Fourchon station (Figure We use a of tide of information from using Mean Time (GMT) and 6 7a,b). min sampling. 24-h period ofacquisition tide information each acquisition, using Time (GMT) 6 min We mark the time offrom the satellite on the plots (red Greenwich vertical lineMean in Figure 7d,e) and and calculate sampling. We mark the acquisition time of the satellite on the plots (red vertical line in Figure 7d,e) and the water level elevation difference between the two acquisition times, which we compare with the calculate the water level elevation between the two acquisition which weto compare InSAR derived water level change.difference We also use the NOAA’s predicted datatimes, for the station, verify with InSAR derived water level change. We also use the NOAA’s for the the InSAR station, that the no extreme events occurred during the InSAR acquisition time that predicted may have data affected to verifyThe thattide no extreme events show occurred during the InSARofacquisition time 11 thatSeptember may have2007–12 affected results. gauge records a water level change 24 cm between the InSAR results. The5tide records show a2009 water level change of 24 cm between 11 September December 2007 and cm gauge between 16 March and 1 May 2009, which are consistent with differences in the fringe pattern the interferograms in Figure 2007–12 December 2007 and 5 cmofbetween 16 March 2009 and 17.May 2009, which are consistent with differences in the fringe pattern of the interferograms in Figure 7.
Remote Sens. 2016, 8, 393 Remote Sens. 2016, 8, 393
9 of 13
9 of 13
Figure 7. 7. (a)(a) ALOS two fringe fringecycles cyclesacross acrossthe thetidal tidal zone mapping Figure ALOSinterferogram interferogramofoftrack track166 166 showing showing two zone mapping thethe tide inundation extent. The white triangle marks the location of the Port Fourchon gauge station. tide inundation extent. The white triangle marks the location of the Port Fourchon gauge station. TheThe white line shows the location of the section described on Figure 7; (b) ALOS 16 March 2009–1 May white line shows the location of the section described on Figure 7; (b) ALOS 16 March 2009–1 2009 interferogram of the same track showing no fringes; (c) Port Fourchon tide gauge data from May 2009 interferogram of the same track showing no fringes; (c) Port Fourchon tide gauge data from 1 September 2007 2007 to 31 to December 2007. 2007. The blue lines show 24-hthe period 1 September 31 December Theand bluegreen and vertical green vertical lines the show 24-h recorded period by recorded the tide gauge for the two dates of dates interferogram (a); (d)(a); Port tide gauge by the tide gauge foracquisition the two acquisition of interferogram (d)Fourchon Port Fourchon tide predicted (solid) and observed (dashed) (dashed) records plotted 24-hour the two gauge predicted (solid) and observed recordsfor plotted for periods 24-hourofperiods ofacquisition the two dates (11 September 2007 in blue 2007 and 12 December in green). The red vertical marks acquisition dates (11 September in blue and 12 2007 December 2007 in green). The redline vertical linethe ALOS acquisition time 4:32 GMT, which 25 cmroughly water level differences between the two marks the ALOS acquisition time 4:32shows GMT, roughly which shows 25 cm water level differences between the acquisitions; (e) A similar tide analysis for thedates two acquisition dates of the(b) acquisitions; (e)two A similar tide gauge analysis forgauge the two acquisition of the interferogram interferogram (b) showing only 5 cm of water level change between the two acquisition times. showing only 5 cm of water level change between the two acquisition times.
The time difference between the ALOS satellite repeat orbit (46 days) and tidal cycle enables The time difference between the ALOS satellite repeat orbit (46 days) and tidal cycle enables InSAR detection of tide-induced water level variations (Figure 7d). However, if the satellite InSAR detection of tide-induced water level variations (Figure 7d). However, if the satellite acquisitions acquisitions and the tidal changes were synchronized, InSAR observations will not be able to detect and the tidal changes were synchronized, InSAR observations will not be able to detect tide-induced tide-induced water level changes occurring at the same tidal amplitude. The tide has both a near daily water occurring the same amplitude. The tide has both near daily h) and (23 level h) andchanges a two-week cycle, at whereas thetidal satellites are sun-synchronous and aacquire data (23 exactly at a two-week cycle, whereas the satellites are sun-synchronous and acquire data exactly at the same time the same time of the day. The phase shift between satellite acquisitions and tide condition results in of thesatellite day. The phase shift between satellite acquisitions and tide condition in satellite observations observations that are collected at different amplitude positionsresults of the tide cycle. The 46-day that are collected at different amplitude positions of the tide cycle. The 46-day interferograms have a interferograms have a small amplitude shift (3 days) producing a minor difference between small amplitude shift (3 days) producing a minor between and, thus, showing acquisitions and, thus, showing no noticeable tidedifference induced changes in acquisitions the interferogram. The 92-days no interferogram noticeable tideobservations induced changes in the interferogram. 92-days interferogram observations are located almost in oppositeThe sides of the tide amplitude cycle, allowingare located almost in opposite sides of the tide amplitude cycle, allowing InSAR to better detect differences
Remote Sens. 2016, 8, 393 Remote Sens. 2016, 8, 393
10 of 13 10 of 13
in the tide inundation level. Within the 3–5 cminundation accuracy of L-band [8,10], theofpatterns InSAR to better detect differences the tide level. Withobservations the 3–5 cm accuracy L-band inferred from[8,10], the fringes, which are in thefrom rangethe of fringes, 7–30 cm,which bearing method and observations the patterns inferred arethe in the rangeuncertainty of 7–30 cm,level bearing hence conclusions madelevel are solid. the method uncertainty and hence conclusions made are solid. inundation extents extents provide a good perspective of tidal inundation through The mapped tidal inundation Louisiana coastal wetlands (Figure 8). The The detected detected tidal tidal inundation inundation extent extent in in the central part of the approximately 65 65 km long long and 15 km wide area of fresh marsh that is being coast shows an area of approximately bytidal tidalcurrents currents(Figures (Figures 8 and Detailed observations the inundation tidal inundation flushed by 8 and 9c).9c). Detailed observations of theoftidal extentextent along along theof coast of Louisiana confirmed that the majority theinundation tidal inundation is being disrupted by the coast Louisiana confirmed that the majority of the of tidal is being disrupted by manman-made and other structures the coast (Figure made canalscanals and other structures alongalong most most of theofcoast (Figure 9). 9).
Figure 8. Wetland classification classification map from Gulf-Wide Information System (G-WIS) (G-WIS) [19] and InSAR detected tidal tidal inundation inundationextents extentsthrough through Louisiana coastal wetlands (black Theframes red frames Louisiana coastal wetlands (black line).line). The red show show the location of detailed observations presented in Figure 9. the location of detailed observations presented in Figure 9.
The The tidal tidal zone zone along along coastal coastal Louisiana Louisiana shows shows to to be be delimited delimited mainly mainly by by dredged dredged canals canals and and levees, levees, leaving leaving narrow narrow areas areas of of uninterrupted uninterrupted tidal tidal inundation. inundation. Some Some areas areas show show disruptions disruptions by by both both levees levees and and man-made man-made canals canals (Figure (Figure 9c) 9c) while while others others are are disrupted disrupted by by canals canals before before reaching reaching the the populated areas where most of the levees are located (Figure 9d). Our results show that at present populated areas where most of the levees are located (Figure 9d). Our results show that at present approximately approximately 40% 40% of of the the Louisiana Louisiana coast coast is is subjected subjected to to uninterrupted uninterrupted tidal tidal inundation. inundation. However, However, with rising sea levels and continuous coastal subsidence the length of the shoreline with with rising sea levels and continuous coastal subsidence the length of the shoreline with uninterrupted uninterrupted tidal will due of toman-made the existence of man-made at short tidal inundation willinundation decrease due to decrease the existence structures at shortstructures distances (1–6 km) distances (1–6 km) inland of the current tidal inundation zone. inland of the current tidal inundation zone. Disruptions along the Louisiana coast are major factors of wetland loss. Anthropogenic Disruptions along the Louisiana coast are major factors of wetland loss. Anthropogenic intervention is transforming coastal Louisiana, changing the natural tide interaction with the coast intervention is transforming coastal Louisiana, changing the natural tide interaction with the coast and and changing the regular sediment deposition with it. The dredged canals that interrupt the water changing the regular sediment deposition with it. The dredged canals that interrupt the water flow flow work as barriers that divide and separate the coastal wetlands. The dissected wetlands become work as barriers that divide and separate the coastal wetlands. The dissected wetlands become more more sensitive to sea level rise of 3.2 ± 0.4 mm/year [31] and ultimately dying due to land subsidence sensitive to sea level rise of 3.2 ˘ 0.4 mm/year [31] and ultimately dying due to land subsidence and and increasing sea level. increasing sea level. This study mainly focuses on the analysis of the interaction between tidal inundation and the coastal wetlands of Louisiana by observing changes in the water level along the coast. However it is important to mention that coastal wetlands are strongly affected and driven by temperature changes, rainfall regimes and sea-level rise [32], and they play a permanent role in the synergy between the coastal wetlands and the tidal flow. Nevertheless, our study shows that man-made structures along
Remote Sens. 2016, 8, 393
11 of 13
the Louisiana coast have changed the regular patterns of tidal inundation through coastal wetlands Remotehence, Sens. 2016, 8, 393 11 of 13 and, have become a dominant element in determining coastal wetland distribution.
Figure 9. Zoom-in maps showing geographical features along the InSAR detected boundary of tidal inundation ofof Louisiana and (d)(d) Timbalier bay,bay, westwest of the inundation extents: extents:(a,b) (a,b)Chenier ChenierPlane; Plane;(c) (c)central centralcoast coast Louisiana and Timbalier of ® world ® world Mississippi River delta. Satellite images source: EsriEsri ArcGIS imagery basemap. imagery basemap. the Mississippi River delta. Satellite images source: ArcGIS
This study mainly focuses on the analysis of the interaction between tidal inundation and the 6. Conclusions coastal wetlands of Louisiana by observing changes in the water level along the coast. However it is The analysis of four Radarsat-1 and twelve ALOS swaths over the Louisiana coast reveals that important to mention that coastal wetlands are strongly affected and driven by temperature changes, tidal inundation is restricted to narrow areas along the coast, where the tide can flow regularly. InSAR rainfall regimes and sea-level rise [32], and they play a permanent role in the synergy between the results show that the Chenier Plain has the narrowest tidal zone, reaching a maximum of 2 km in some coastal wetlands and the tidal flow. Nevertheless, our study shows that man-made structures along parts, and the Mississippi River delta has wider zones reaching up to 15 km. InSAR technique showed the Louisiana coast have changed the regular patterns of tidal inundation through coastal wetlands to be an effective tool to map and measure the extents and amplitude of the tide inundation zone and, hence, have become a dominant element in determining coastal wetland distribution. through Louisiana coastal wetlands, as well as to detect areas of tidal flow disruptions along the coast. Detailed observations of the produced interferograms indicate that man-made canals and levees are 6. Conclusions the main sources of tide disruptions along the entire coast. The analysis ofbetween four Radarsat-1 and twelve ALOS swaths overand the tide Louisiana reveals that A comparison InSAR detected water level changes gaugecoast records indicate tidalinterferometric inundation is restricted to narrow areaswetlands along thehighly coast, where the flow regularly. that observations in coastal depend ontide thecan tide-induced waterInSAR level results show that the Chenier Plain has the narrowest tidal zone, reaching a maximum of 2 km in changes observed between satellite acquisitions. In the case of the coast of Louisiana, the 92-day ALOS some parts, and the Mississippi River delta has wider zones reaching up to 15 km. InSAR technique interferograms observe larger tide-induced water level changes than the 46-day ones, providing clearer showed to beofan tool to map and measure theour extents amplitude ofLouisiana the tide inundation information theeffective tidal inundation extents. Although studyand is limited to the coast, the zone through Louisiana coastal wetlands, as well as to detect areas of tidal flow disruptions along use of this technique can be expanded to detect tidal inundation interactions in similar coastal zones. the coast. Detailed observations of the produced interferograms indicate that man-made canals and The implementation of InSAR as a monitoring tool to detect tidal inundation propagation can provide levees are the main sources of tide disruptions along the entire coast. A comparison between InSAR detected water level changes and tide gauge records indicate that interferometric observations in coastal wetlands highly depend on the tide-induced water level changes observed between satellite acquisitions. In the case of the coast of Louisiana, the 92-day ALOS interferograms observe larger tide-induced water level changes than the 46-day ones,
Remote Sens. 2016, 8, 393
12 of 13
useful constrains for coastal tidal inundation models, and can be used to detect and track sources of tidal flow disruptions. Supplementary Materials: The following are available online at www.mdpi.com/2072-4292/8/5/393/s1, Table S1: Satellite specifications. Figure S1: ALOS 46-92 day interferograms from track 166. Acknowledgments: Authors thank the three anonymous reviewers for their very helpful comments on the manuscript. We would like to thank Sang-Hoon Hong and Heresh Fattahi for their help processing the InSAR data. We also acknowledge help of the Japan Aerospace Exploration Agency (JAXA), the Canadian Space Agency (CSA), the Alaska Satellite Facility (ASF) and the National Oceanic and Atmospheric Administration (NOAA) for providing access to the satellite images and tide gauge station information. We thank the Fulbright and CONACYT grants for their support of the first author. The research was supported by the National Science Foundation through Research program under Grant No. DEB-1237517. Author Contributions: Talib Oliver was responsible for InSAR and tide-gauge data processing. Both authors contributed to the data interpretation and the writing of the manuscript. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3.
4. 5. 6. 7.
8. 9. 10. 11. 12. 13.
14. 15. 16. 17.
Gallant, A.L. The challenges of remote monitoring of wetlands. Remote Sens. 2015, 7, 10938–10950. [CrossRef] Couvillion, B.R. Land Area Change in Coastal Louisiana (1932 to 2010); US Department of the Interior, US Geological Survey: Reston, VA, USA, 2011. Penland, S. Geomorphic classification of coastal land loss between 1932 and 1990 in the Mississippi River Delta Plain, Southeastern Louisiana. USGS Open-File Report 00-417; US Geological Survey Woods Hole: Barnstable County, MA, USA, 2000. Steyer, G.D.; Couvillion, B.R.; Barras, J.A. Monitoring vegetation response to episodic disturbance events by using multitemporal vegetation indices. J. Coast. Res. 2013, 63, 118–130. [CrossRef] Barras, J. Historical and Projected Coastal Louisiana Land Changes: 1978–2050; United States Geological Survey: Reston, VA, USA, 2003. Ramsey, E.; Rangoonwala, A.; Bannister, T. Coastal flood inundation monitoring with satellite C-band and L-band synthetic aperture radar data. JAWRA J. Am. Water Resour. Assoc. 2013, 49, 1239–1260. [CrossRef] Moffett, K.B. Multiple stable states and catastrophic shifts in coastal wetlands: Progress, challenges, and opportunities in validating theory using remote sensing and other methods. Remote Sens. 2015, 7, 10184–10226. [CrossRef] Wdowinski, S. Space-based detection of wetlands’ surface water level changes from L-band SAR interferometry. Remote. Sens. Environ. 2008, 112, 681–696. [CrossRef] Kim, J.-W. Integrated analysis of PALSAR/Radarsat-1 InSAR and ENVISAT altimeter data for mapping of absolute water level changes in Louisiana wetlands. Remote Sens. Environ. 2009, 113, 2356–2365. [CrossRef] Sang-Hoon, H.; Wdowinski, S.; Sang-Wan, K. Evaluation of TerraSAR-X Observations for Wetland InSAR Application. IEEE Trans. Geosci. Remote Sens. 2010, 48, 864–873. [CrossRef] Kim, S.-W. Interferometric Coherence Analysis of the Everglades Wetlands, South Florida. IEEE Trans. Geosci. Remote Sens. 2013, 51, 5210–5224. [CrossRef] Pritchard, M.E. InSAR, a tool for measuring Earth’s surface deformation. Phys. Today 2006, 59, 68. [CrossRef] Lu, Z.; Kwoun, O.-I. Radarsat-1 and ERS InSAR analysis over southeastern coastal Louisiana: Implications for mapping water-level changes beneath swamp forests. IEEE Trans. Geosci. Remote Sens. 2008, 46, 2167–2184. [CrossRef] Elijah, R., III; Lu, Z.; Rangoonwala, A.; Rykhus, R. Multiple baseline radar interferometry applied to coastal land cover classification and change analyses. GISci. Remote Sens. 2006, 43, 283–309. Wdowinski, S. Remote-sensing monitoring of tide propagation through coastal wetlands. Oceanography 2013, 26, 64. [CrossRef] Boesch, D.F.; Josselyn, M.N.; Mehta, A.J.; Morris, J.T.; Nuttle, W.K.; Simenstad, C.A.; Swift, D.J. Scientific assessment of coastal wetland loss, restoration and management in Louisiana. J. Coast. Res. 1994, 20, i-103. Owen, D.E. Geology of the Chenier Plain of Cameron Parish, southwestern Louisiana. Field Guides 2008, 14, 27–38.
Remote Sens. 2016, 8, 393
18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.
13 of 13
Sasser, C.E. Vegetation types in coastal Louisiana in 2013. U.S. Geological Survey Scientific Investigations Map 3290; US Department of the Interior, US Geological Survey: Reston, VA, USA, 2013. Minerals Management Service. G-WIS Environmental Sensitivity Index Datasets; Minerals Management Service: New Orleans, LA, USA, 2002. MDA Federal. Landsat Geocover 2000/ETM+ Edition Mosaics. Available online: http://glcf.umd.edu/ data/mosaic/ (accesed on 5 July 2015). Bernier, J. Trends and Causes of Historical Wetland Loss in Coastal Louisiana. Available online: http://pubs.usgs.gov/fs/2013/3017/ (accessed on 4 May 2016). Dixon, T.H. Space geodesy: Subsidence and flooding in New Orleans. Nature 2006, 441, 587–588. [CrossRef] [PubMed] Morris, J.T. Responses of coastal wetlands to rising sea level. Ecology 2002, 83, 2869–2877. [CrossRef] University of Alaska Fairbanks. The National Aeronautics and Space Administration. Available online: https://ursa.asfdaac.alaska.edu/ (accessed on 5 February 2015). Richards, J.; Woodgate, P.; Skidmore, A. An explanation of enhanced radar backscattering from flooded forests. Int. J. Remote Sens. 1987, 8, 1093–1100. [CrossRef] Wdowinski, S.; Hong, S.-H. 7 Wetland InSAR. In Remote Sensing of Wetlands: Applications and Advances; CRC Press: Boca Raton, FL, USA, 2015. Rosen, P.; Fielding, E. Repeat Orbit Interferometry Package (ROI_PAC). Available online: http://roipac.org/ (accessed on 6 May 2016). Hanssen, R.F. Radar Interferometry: Data Interpretation and Error Analysis; Springer Science & Business Media: Berlin, Germany, 2001. The National Oceanic and Atmospheric Administration. Center for Operational Oceanographic Products and Services. Available online: http://tidesandcurrents.noaa.gov/ (accessed on 5 May 2016). Penland, S.; Suter, J.R. The geomorphology of the Mississippi River Chenier plain. Mar. Geol. 1989, 90, 231–258. [CrossRef] Church, J.A.; White, N.J. Sea-level rise from the late 19th to the early 21st century. Surv. Geophys. 2011, 32, 585–602. [CrossRef] Osland, M.J. Beyond just sea-level rise: Considering macroclimatic drivers within coastal wetland vulnerability assessments to climate change. Glob. Chang. Biol. 2016, 22, 1–11. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
