Auxiliary material for Paper 2011GL047174

Formation dynamics of subsurface hydrocarbon intrusions following the Deepwater 
Horizon blowout

Scott A. Socolofsky 
Zachry Department of Civil Engineering, Texas A&M University, 
College Station, Texas, USA

E. Eric Adams 
Ralph M. Parsons Laboratory, Massachusetts Institute of Technology, 
Cambridge, Massachusetts, USA

Christopher R. Sherwood
U.S. Geological Survey, Woods Hole, Massachusetts, USA



Socolofsky, S. A., E. E. Adams, and C. R. Sherwood (2011), Formation dynamics of 
subsurface hydrocarbon intrusions following the Deepwater Horizon blowout, 
Geophys. Res. Lett., 38, LXXXXX, doi:10.1029/2011GL047174.

Introduction

This auxiliary material describes our methods for processing CTD (conductivity, 
temperature, depth) profiles from the Gulf of Mexico, making estimates of 
kinematic buoyancy flux B, and using these to estimate trap height hT and 
equivalent buoyancy frequency NE.

CTD PROFILES

We examined all of the ship data publically available (as of 12 December 2010) 
on the NOAA/NODC website:

http://www.nodc.noaa.gov/General/DeepwaterHorizon/oceanprofile.html last 
accessed 11 February 2011  

We found 267 CTD profiles that were deeper than 700 m and made within 7 km of 
the wellsite between early May and mid July 2010. We did not do any quality 
control or editing of the profile data. We calculated potential temperature and 
potential density from recorded temperature and salinity using Matlab® routines 
from the CSIRO (Commonwealth Scientific and Industrial Research Organisation) 
SeaWater Toolbox version 3.3:

http://www.cmar.csiro.au/datacentre/ext_docs/seawater.htm; last accessed 1 
February 2011  

Fluorescence was measured in many of the profiles with a Wetlabs ECO-AFL/FL and 
recorded in units of chromomorphic dissolved organic matter (CDOM; mg/m^3).  
There are 92 profiles with "excess" fluorescence.  Excess fluorescence was 
determined by constructing a linear background profile determined by the mean 
CDOM level above the intrusions (between depths of 800 - 850 m) and the mean 
CDOM level in the bottom 50 m of the CTD cast. CDOM measurements more than 4 
standard deviations above this background were deemed "excess".  This method 
worked well for defining excursions in the CDOM profiles except in a few cases 
where the CTD cast terminated in a CDOM peak; they are not included here.  The 
range in excess values was typically only 5 - 20 mg/kg^3, but Figure 4 in the 
journal paper is dominated by several profiles where excess CDOM was much 
larger.

TIME-DEPENDENT KINEMATIC BUOYANCY FLUX

The initial kinematic buoyancy flux Bo depends on the volume flow rate of the 
oil and gas.  We used the estimates published by the Flow Rate Technical Group 
(FRTG; McNutt et al., 2011).  Estimated flow of oil from the well declined 
linearly from ~62,000 bbls/day on 22 April to ~53,000 bbls/day on 15 July 
(Figure S1; top panel), with an uncertainty of ±10%.  The FRTG estimated that 
flow rates were about 4% higher than this trend between 3 June (when the riser 
was cut) and 12 July (when the new capping stack was closed).  An average of 
~2,430 bbls/day of oil was recovered using the riser insertion tube tool (RITT) 
between 16 May and 24 May, and an average of ~18,700 bbls/day of oil was 
recovered using the "top hat" between 8 June and 16 July. We subtracted the 
daily recovery volumes to estimate the net flow of oil to the subsurface plume.  
We assumed a constant oil:gas ratio at the well head of 1:0.79 throughout the 
spill.  Accordingly, the mean in situ release rates of oil and gas were 0.09 
m^3/s and 0.07 m^3/s, respectively. The time series of Bo based on net oil flow 
rates is shown in Figure S1, bottom panel and listed in Data Set S1. The gray 
band incorporates the 10% uncertainty suggested by the FRTG.  Before the riser 
was cut, we have assumed the lower estimate is 50% of the total flux because the 
flow may have been partitioned between two separate sources. Mean combined 
(oil+gas) Bo during the event was 0.78 m^4/s^3.

TRAP HEIGHT AND EQUIVALENT BUOYANCY FREQUENCY

We estimated quadratic density profiles between depths of 800 and 1510 m for 
profiles within 7 km of the wellhead for 122 profiles deeper than 1510 m (Data 
set 1).  We performed least-squares linear fits to the quadratic form rhof(z) = 
rho0 + bz^2, where rhof(z) is the fitted density profile (kg/m^3), rho0 is the 
intercept (at 1510 m, where z = 0), b is the slope of the fit (kg/m^5), and z is 
the elevation above the nominal depth of the wellhead (1510 m).  Differentiating 
the quadratic fit, and using the definition of buoyancy frequency N^2 = -g/rho0 
drho/dz, we determined the slope a of a linear buoyancy frequency profile (Eqn. 
3 in the journal paper) is a = -2gbz/rho0 m^(-1) s^(-2). Using the Deepwater 
Horizon values (Table 2 in the journal paper) for slip velocity (us), initial 
buoyancy flux B0 estimated from flow rates (Figure S1), and an initial estimate 
of buoyancy frequency N = 0.002 s-1, we evaluated the constant c (Eqn. 7 in the 
journal paper) and used it, along with a and m=1 to determine trap height hT 
(Eqn. 8 in the journal paper).  The associated equilibrium buoyancy frequency NE 
was then determined with Eqn. 5 in the journal paper. Trap height estimates from 
these calculations are shown as black dots in Figure 4 of the journal paper.  
Data set 1 lists the associated values for NE and values for the fits to density 
profiles (a, rho0, b, and squared regression coefficient r^2).  These fits were 
remarkably good (median r^2 was 0.9959 and ranged from 0.9925 to 0.9986) and 
fairly constant (slope parameter b varied between -5.49x10-7 and -4.14x10-7 
kg/m^5) suggesting that the density gradient was quite linear and did not vary 
much between early May and mid July. This is reflected in the relatively 
constant calculated trap heights hT (median 366, range 352 - 381 m) and 
equivalent buoyancy frequencies NE (median 0.00156, range 0.00146 - 0.00164 s^(-
1)).

1. 2011gl047174-fs01.eps  
Figure S1. Time line of oil flow rates and initial kinematic buoyancy fluxes. 
Upper panel: oil flow rates estimated by the Flow Rate Technical Group (FRTG; 
McNutt et al., 2011; red dashed line) and net flow rate into the plume after 
subtracting oil recovered by the riser insertion tube tool (RITT) and the "top-
hat" (black line). Lower panel: estimated time series of combined (oil+gas) 
initial buoyancy flux, with uncertainty estimates. These values were used to 
estimate trap height hT in Figure 4 of the journal article.

2. 2011gl047174-ds01.txt 
Data Set S1. Time series of Bo based on net oil flow rates. In Comma separated 
values format.

2.1 Col. 1: Data file name
2.2 Col. 2: Cast number 
2.3 Col. 3: Station Name
2.4 Col. 4: Date
2.5 Col. 5: Longitude (decimal degrees)
2.6 Col. 6: Latitude (decimal degrees)
2.7 Col. 7: Distance from wellsite (km)
2.8 Col. 8: Compass bearing from wellsite to station (degrees)
2.9 Col. 9: Maximum depth (m)
2.10 Col. 10: Initial kinetic buoyancy flux Bo (m^4/s^3)
2.11 Col. 11: Trap height hT (m)
2.12 Col. 12: Equivalent buoyancy frequency NE (s^(-1))
2.13 Col. 13: Coefficient a (m^(-1) s^(-2))
2.14 Col. 14: Density offset at z=0 from fit (kg/m^3)
2.15 Col. 15: Slope b (kg/m^5) from fit
2.16 Col. 16: Correlation coefficient r^2

Disclaimer. Use of trade, product, or firm names is for descriptive
purposes only and does not imply endorsement by the U. S. Government.