Characterizing Tidal Inundation of Wetlands in the Murderkill Estuary (Kent County, DE)

Characterizing Tidal Inundation of Wetlands in the Murderkill Estuary (Kent County, DE) Tom McKenna Delaware Geological Survey University of Delaware Thermal Imaging Can temperature be used as an indicator of inundation? LiDAR Surface Elevation Tidal Inundation January 14, 2009 Supported by Kent County & Delaware DNREC

What causes low DO in the tidal Murderkill River? Issue: low DO (EPA impaired waters ) Question: appropriate TMDLs for nutrients? (EPA pollution control strategy ) river / marsh biogeochemical i interaction? ti Webbs Marsh July 3, 2008 low tide WWTP extensive salt marsh Waste Water Treatment Plant? NJ DE Bay DE

McGinnis Pond Cou rse y Pond Andrews Lake McColley Pond De laware Ba y Primary Research Question What processes, sources, and sinks control low DO in the tidal Murderkill River? Numerical Water-Quality Modeling site-specificspecific conceptual models and input data parameterizations of processes that are difficult to model explicitly and/or require information not easily obtainable tidal inundation of marshes biogeochemical cycling in marshes nutrient fluxes in/out marshes primary productivity sediment oxygen demand WWTP

Tidal inundation of marshes The flow of water on salt marsh platforms is still poorly characterized. New conceptual models are being developed. Challenges: microtopography groundwater / surface-water interaction anthropogenic alteration (since late 1600s) inundation / vegetation relationship very dynamic system dense vegetation overland flow in estuary-scalescale models marsh platform upland primary tidal channel secondary tidal channel Discrepancies in tidal phase and elevation in a numerical model can be accommodated by the modeling calibration process but can severely limit the explanatory power and predictive capabilities of the model. (French, 2003)

Methods for determining inundation Hydrodynamic Model (Overland Flow) Tide Gages Elevation Survey Vegetation Survey Comment Model based on distribution of vegetation communities (indirect) No Few Yes Yes assumes correct No No No Yes conceptual model Model based on tide and ground-surface elevation (direct) Yes Many ground-based Yes Cadillac Yes/No Many/Few ground-based or LiDAR Yes/No No Few LiDAR No Temperature

common x,y,z datum apples to apples Components of elevation marsh surface (LiDAR) very low relief minimize error and remove bias instrument locations (e.g. tide level) preliminary inundation model highly parameterized data loggers Inundation Study water level, temperature, and salinity time-series to document inundation and develop & test models air temperature & pressure, solar radiation temperature on marsh platform (air/water) Can we use temperature as an indicator of inundation? thermal imaging Can we use temperature as an indicator of inundation?

Model for tidal inundation Preliminary model makes VERY simple (and wrong) assumption of Instantaneous inundation of entire marsh based on water elevation at Bowers Beach. How wrong? Working towards more complex representation: tidal propagation: function for instantaneous tide level for specified reaches of Murderkill River using data from tide gages at Frederica and Bowers Beach TL r = f(xr,t) assign TL r to a subset of marshes along the reach marsh subset inundated based on TL r and set of spatial parameters representing (raster layers): marsh elevation attenuation/amplification of tidal amplitude phase shift of tide based on distance from river/creek/ditch

Testing model of tidal inundation Direct measurement: water level, temperature, and salinity in tidal creeks / ditches using data loggers (time and instrument intensive). Marsh platform instrumentation? ALTERNATIVE: Can observations of environmental temperature be used to capture spatio-temporal temporal dynamics of inundation (less expensive, more representative of spatial heterogeneity)? Water flow and heat transfer are highly-coupled processes. 7:19 DST 105 Temperature loggers on marsh platform Environmental Thermography

Establishing a common vertical datum low relief on marsh platforms so minimizing elevation error and bias is critical asset source datum vertical type datum LiDAR monuments USGS DELDOT ellipsoidal ellipsoidal NAD83 NAD83 monuments NGS ellipsoidal NAD83 monuments NGS orthometric NAVD88 tide gages USGS orthometric NGVD29 tide prediction NOAA tidal MLLW Geoid model: GEOID03 / VDATUM software (NOAA) convert between ellipsoid, orthometric, and tidal datums Conducted least squares adjusted GPS-RTK survey 8 reference monuments; minimally constrained

Survey corrections Elevation adjustments of up to 55 cm! Significant for accurate modeling of marsh platform inundation and hydrodynamics in river. Correction (m) Location reference elevation value NGVD29 to NAVD88 least square adjust Total (m) USGS Bowers gage -0.34-0.24 +0.03-0.55 USGS Webbs gage -0.34-0.24 +0.03-0.55 USGS Frederica gage 0.00-0.24 +0.08-0.16 N Bowers ref. monument +0.03 +0.03 Barretts Ch. ref. monument +0.08 +0.08 LiDAR ref. monument +0.08 +0.08

ELEVATION MAPPING LIDAR survey by USGS / NASA Mark Nardi & Wayne Wright 2008; low spring tide; leaf-off EAARL Experimental Advanced Airborne Ranging g LiDAR

Digital Elevation Model LiDAR survey scale optimized for marsh topography Murderkill River salt marsh uplands LiDAR data collected for other purposes by commercial LiDAR vendors is not optimal for mapping marsh topography. Delaware Bay

Cumulative Probability of Tide Being Below an Absolute Elevation 3 gages (USGS) Bowers Frederica Webbs Slough tidal datums shown for Bowers (VDatum) Preliminary model makes VERY simple (and wrong) assumption of Instantaneous inundation based on water elevation at Bowers Beach But how wrong? Fraction of time wat ter level be elow given elevation -1.5 MLLW MLW MTL MH HW 85% of time below MHW 75th % MH HHW 45% of time below MTL 2% of time below MLW Elevation (m) median 25th %

Digital Elevation Model LiDAR Delaware Bay Murderkill River uplands salt marsh

elevation < mean tide level 002mNAVD88-0.02 m NAVD88 LSA using tide levels from Bowers Beach gage

elevation < 0.4 m 04mNAVD88 0.4 m NAVD88 LSA using tide levels from Bowers Beach gage

elevation < mean high water 06mNAVD88 0.6 m NAVD88 LSA using tide levels from Bowers Beach gage

elevation < mean highest high water 075mNAVD88 0.75 m NAVD88 LSA using tide levels from Bowers Beach gage

elevation < 0.9 m 09mNAVD88 0.9 LSA (> MHHW) using tide levels from Bowers Beach gage

Data Loggers in Webbs Marsh ~ 1 year at 6-minute sampling interval temperature temperature, salinity, level temperature, salinity, level, velocity, ph, DO Murderkill R. 3 2 1 WS Murderkill R

Likely Temperature Signals from Inundation? season air temperature water day nite spring warm cool warm/cool summer warm cool warm fall cool cool warm/cool winter cool cold cool

Temperature signals of inundation at 3 sites on the marsh platform (over 7 days) inundation progresses through time from site 1 to site 3 Jun 1, 2008 Jun 8, 2008 o C 1 night day start of inundation at site 1 start of inundation at site 3 o C o C 2 3 o C Water temperature @ Webbs Slough (WS) air temp. temperature difference (site 1 site WS) m Webbs Slough (WS) tide

Phase lag of inundation relative to tide level at Webbs Slough area minutes secondary ditches 10-20 Spartina alterniflora 20-40 near ditches Spartina alterniflora 60-90 away from ditches Spartina patens 60-90

MULTI-SPECTRAL IMAGING Reflected Energy Emitted Energy Thermal Bands UV Visible 0.4µm NIR SWIR 1.0µm 1.7µm MWIR 3.0µm LWIR 5.0µm thi study this t d 8.0µm 14.0µm

Environmental Thermography (Thermal Imaging of the Environment) measure apparent temperature of remote surfaces located centimeters to kilometers away absolute temperature can be calculated emissivity of a material (e.g. water, sediment, vegetation) geometry of the observing system can be used as a quick investigative tool for detailed lab / field studies of dynamic processes natural, engineered, industrial

Imaging g Platforms

lift platform Temperature Time Series of Tidal Mixing in Webbs Slough (May 15, 2008) 12 C 18 visual Cooler water from the Murderkill River flows into Webbs Slough during flood tide and back out during ebb tide 5:34 DST 5:49 DST 6:04 DST 6:19 DST 0 min flood tide sunrise 15 30 45 6:34 DST 6:49 DST 7:04 DST 7:19 DST high ebb 75 tide tide 60 75 90 105 7:34 DST 7:49 DST 8:04 DST 8:19 DST 120 135 150 165 min

Temperature Time Series Showing Inundation of Marsh Platform A flood tide 8:44 PM EST B 816-817 warmer water (gray blue) on colder (red) marsh surface visual 12/13/2008 B A thermal imager location B high tide 9:04 PM EST A 882-885 B ebb tide 9:42 PM A 1062-1064 July 31, 2008 early evening

Image of Marsh Platform Inundation warmer water flowing onto cold marsh surface Webbs Marsh December 13, 2008 high spring tide cold morning (subfreezing) helicopter platform inundated platform is warm (orange) B A B A channels are hot (yellow)

Differentiating Sources of Water Visual Image Thermal Image cold water (black) ponded on marsh platform after recent rain event B B A A warm water (white) in Brockonbridge Gut, Webbs Slough, and ditches 12/13/2008 9:56 AM high tide

Conclusions Salt-marsh elevations in upper Murderkill Estuary (Frederica) are lower than lower estuary (Bowers Beach) Common vertical datum is critical for quantifying tidal inundation using in-situ instrumentation. Environmental temperature can be used to capture spatio-temporal dynamics of inundation. inexpensive in-situ data loggers environmental thermography