Elkhorn Slough National Estuarine Research Reserve: Bank erosion and rapid mudflat biomonitoring

Scientists from the Elkhorn Slough National Estuarine Research Reserve and the Monterey Bay National Marine Sanctuary conducted surveys in Elkhorn Slough, California, USA every other year from 2001-2023 to assess bank erosion, mudflat erosion, and to conduct rapid biomonitoring surveys of the mudflats.  There are thus three components associated with this dataset:

Bank Erosion: The purpose of this monitoring program was to track erosion rates of banks along the main Elkhorn Slough channel and at various sites on the Reserve, as a component of overall habitat change analysis (bathymetry data, remote sensing of marshes and tidal creeks) shedding light on tidal scour.  Erosion rates can be correlated to other measured variables, such as cliff height or crab hole density.

Mudflat Erosion: The purpose of this monitoring program was to track erosion or deposition rates in intertidal mudflats at sites throughout the estuary.

Rapid Mudflat Biomonitoring: One purpose of this program was to detect new invasions, potentially early enough to attempt eradication, and to track changes in distribution and abundance of a set of conspicuous existing invaders.  We also were interested in detecting major changes in bivalve populations or other native ecological dominants.

Methods
Bank Erosion Monitoring

At each station, there was an upstream and downstream PVC stake 10 m apart, parallel to the shore.  We ran a transect tape perpendicular measured the distance from the stake to the bank edge and to the vegetation edge (the most seaward live plant in transect).  While the primary interest was bank erosion, at stations with low or multiple cliffs, it wasn’t always clear which to use as the bank edge, so the measurement to vegetation edge is more repeatable. When there was no clear bank edge at all, we used distance to vegetation edge as the distance to bank edge measurement. As necessary we moved stakes backwards from edge to prevent their loss, and add this amount to datasheet to adjust calculations.  We averaged the data from the two stakes per site (useful because if one stake is missing, we still had a measurement for that site).  From these data we calculated annualized bank erosion rates for different Slough regions. We also assessed a series of other parameters, hole density, bank height and undercut.  

In 2013, we added shorter stakes 1 m (or closer on very narrow banks, specified on datasheet if different than 1 m) behind original tall PVC to allow us to have a repeatable angle for bank measurements, and to make it more likely that at least one of the stakes would still be present two years later, so we wouldn’t lose the time series as we’d done previously if a stake was missing. From 2013 forward, we conducted all measurements from the most landward shorter stake (so distance from PVC to shorter stake, usually 100 cm, was subtracted from all measurements from 2013 forward to make them equivalent to earlier ones).

Because of some evolution of methods over time, and because of missing data from some years and stakes, the raw data for this monitoring is complicated and not user-friendly.  We have archived the raw data (“D erosion raw”) in an Excel file for completeness, but strongly recommend that users used the curated data (the CSV) for statistical analyses. The curated dataset uses the average value of the two stakes per station (providing a more consistent time series because in many early years we had missing stakes and lost time series), and omits stations with artificial banks and sites with muted tidal exchange.  It also omits the plant type and cover measurements taken in early monitoring years, which proved to be unreliable due to variation in placement of the quadrats.

Mudflat erosion

After 2013, we also added a new element, mudflat erosion/deposition, evaluated by measuring how many cm of a short PVC stake were emergent above the mudflat surface, and quantifying changes in this over time.  The short PVC stakes placed at +1.5 ft above MLLW in the intertidal, in the transect line from each bank stake. 

Rapid Mudflat Biomonitoring

We carried out the monitoring at all the bank erosion stations (mudflats).  In the 10 m long rectangle in between the bank stakes and the mudflat stakes (located at +1.5 ft above MLLW), we assessed the mudflat. While the search area was always 10 m long (parallel to shore), it varied in width depending on slope, yielding a search area typically 10-100 square meters.  Before 2013 we estimated search area; after 2013 it was marked by the low PVC stakes placed at +1.5 ft above MLLW and was measured; it remained the same from then on. At each station, we searched for a targeted list of conspicuous epifaunal organisms, mostly invaders but also some key natives. Distinctive burrow openings were included (gapers, ghost shrimp).  Each was given an abundance value (0-4, see Section 8 for scoring).  If other organisms (greater than about 2 cm in size) were also abundant (scores 3-4), we identified them to whatever level we could. Relative abundance was calculated as a function of available habitat, separately for organisms on mudflat (most of area) vs. organisms on hard substrates (a tiny percentage of the search area).  E.g., we assessed what percentage of the only two rocks present at the site had barnacles, rather than what percentage of the mudflat had barnacles. 

Over time, we added new taxa that we hadn’t previously searched for, such as Ulva and the holes of crabs.  For such species, their abundance in early years is listed as NA rather than zero, and should be treated differently than zeros.  For a time series of any particular species, one should look only at those years in which the species was assessed (has numerical data, not NAs).

The field datasheets with exact definitions used and field instructions are archived as tabs on the Excel database for this monitoring dataset (“M field erosion datasheet”, “M field bio datasheet”, “M field instructions”).

Sites

Monitoring occurred at about 20 stations each year. Those along the main channel were established by Chris Malzone in the 1990s, and numbering is his.  New stations were added on ESNERR in 2001.  At each station, there was an upstream and downstream stake, 10 m apart on the bank.  These were often lost in early years (eroded away), which meant rate of change in distance from stake to edge could not be calculated.  Later, a second stake 1 m back from first was added, making it easier to replace one of the stakes if lost.  Mudflat stakes were added in 2013 to delineate the biomonitoring search area and to monitor mudflat erosion.  

Over time, some stations were dropped.  This applied to various ones on eroding levees, which had eroded so badly that there was no space left for measuring bank erosion.  

Stations are described and coordinates provided in the “M Station Info” tab of the Excel database.

Bank Erosion Monitoring

The curated CSV includes only those stations on natural banks with full tidal exchange.  Stations on artificial banks were dropped because levees entirely eroded away, because measurements were unreliable (very variable from year to year in unrealistic ways), and because the main focus of this monitoring is on loss of natural salt marsh banks.  Stations with muted tidal exchange were dropped for the latter two reasons. 

Sites with low cliffs (indicated in column in data) can optionally be omitted from analysis of bank erosion rates, since these sites have less reliable measurements (they usually have multiple low banks so there is some choice as to which to measure to) and since bank erosion studies typically focus on higher banks.

Mudflat erosion

The curated CSV includes all stations except 2 (no stakes installed) and stations on the Reserve that were not monitored for long after stakes were installed (61, 80-81).  

Mudflat Biomonitoring

The curated CSV includes all stations.

Data collection period

Bank erosion monitoring occurred from 2001-2023, mudflat erosion monitoring from 2015-2023, and rapid mudflat biomonitoring from 2005-2023. Monitoring was typically conducted on two Fall season low tides, in odd years, though a few even were done early on (2002 and 2004).  On one day, walk-in stations were assessed by three teams working separately; on the other day, boat-based stations were jointly assessed by one large team.

Survey dates and team members are listed on the “M survey dates teams” metadata tab in the Excel database.

Explanation of variables

The field datasheets with exact definitions used and field instructions are archived as tabs on the Excel database for this monitoring dataset (“M field erosion datasheet”, “M field bio datasheet”, “M field instructions”). Only a brief overview is provided below because detailed metadata are provided in these tabs. All clean datasets include the region, subregion, area and subarea using standard placenames designated for all Elkhorn Slough datasets to allow for cross referencings.

Bank Erosion Monitoring

The core data collected are distance from marker to bank and vegetation edge, which are converted into annualized rates of change averaged between the two stakes per station in the clean dataset for analyses. Annualized rates are much more reliable and user-friendly than raw distances, because these changed as stakes were replaced and because monitoring intervals varied. Note that the annualized rate of change covers the two-year period preceding each survey year (e.g., the annualized rate listed for 2023 describes what happened from Fall 2021-Fall 2023); this is important to consider if correlating rates to environmental variables such as ENSO index (which should thus also be averaged for the same two year period). Supplemental data collected include cliff height, undercut, distance between vegetation and bank edge, and large and small hole in a 0.25 m-2 quadrat.  All measurements are in cm (or for rates, cm/year).  

Key variables 

Annualized_Change_Distance_To_Bank_Edge: if negative, erosion rate; if positive, would be progradation; cm/year

Annualized_Change_Distance_Veg_Edge: if negative, landward marsh retreat; if positive, seaward expansion; cm/year

Annualized_Change_Bank_To_Veg_Edge: change in distance between bank edge and vegetation edge; negative numbers mean distance is getting smaller; positive means distance is getting larger (vegetation retreat); cm/year

Mudflat Erosion Monitoring

Erosion/deposition rates were calculated by comparing how much of each marker PVC in mudflat was protruding in consecutive measurement years and converting this to annualized rates of change, averaged across the two stakes per station. Measurements are in cm/year.

Key  variable:

Annualized_Change_Sediment_At_Mudflat_Stake: if negative, erosion rate; if positive, deposition rate; cm/year

Rapid Mudflat Biomonitoring

The data consist of a relative abundance estimate for live species.

0=absent

1=trace presence: present on <1% of available substrate

2=rare: on 1-10% of available substrate

3= abundant: on 10-50% available substrate  

4= very abundant: on >50% available substrate

For dead bivalves, the abundance is a count of valves (individual shells – so a complete clam shell would have a value of 2, while a single half-shell would have a value of 1).