Geology of California's heavy oil fields

Paul's first assignment for Ethan is to understand the sands at Tamarisk Hill. 'In a steamflood, geology decides where the heat goes,' he says. 'Two patterns that look identical on the map can have steam-oil ratios that differ by a factor of two because one of them has a clay layer that steers the steam.' Ethan spends his first weeks with the geologist, Grace Nakoochee, going through logs, old core descriptions and the temperature surveys from the observation wells.

This lesson describes the deposits the asset produces from: the shallow, unconsolidated sands of the San Joaquin Valley's heavy oil fields, with the Kern River Formation as the main example, how they were laid down and why they hold heavy oil. It then covers diatomite reservoirs, the clay and shale layers that steer steam, the limits on injection pressure, and how core and logs define the pay a steamflood can heat.

Shallow sands at the edge of a basin

The San Joaquin Valley is a deep sedimentary basin between the Sierra Nevada and the Coast Ranges. Organic-rich marine shales, including the Monterey Formation, generated oil that migrated toward the basin margins and up into younger, shallow sands. On the eastern side of the valley near Bakersfield, the Kern River Formation consists of Pliocene and Pleistocene sands and gravels shed from the Sierra Nevada by rivers and alluvial fans. Oil reached these sands, which are only a few hundred to about thirteen hundred feet below the surface at Kern River, and was biodegraded into heavy oil. On the western side of the valley, fields such as Midway-Sunset produce heavy oil from sands in folded structures along the edge of the Coast Ranges.

The Kern River sands were laid down by braided rivers and alluvial fans, so they are coarse, poorly sorted in places and interbedded with silts and clays from floodplains and abandoned channels. Individual sand bodies are a few feet to several tens of feet thick, and the clay layers between them range from thin, discontinuous drapes to beds that extend across many patterns. The sands are unconsolidated: they have never been cemented, so core falls apart if not frozen or carefully preserved, and the wells need sand control. Porosity is typically around 30 percent and permeability is commonly one to several darcies.

Why the sands are so rich

The Kern River sands were never deeply buried or cemented, so they keep very high porosity and permeability. Oil saturations in the best sands were originally around 50 to 60 percent, lower than in the Athabasca oil sands but in rock that conducts fluids extremely well. The oil gravity at Kern River is about 10 to 15 API, and its viscosity at the original reservoir temperature of roughly 30 °C (90 °F) is in the thousands of centipoise, too viscous for useful cold production but light enough to flow readily when heated to a few hundred degrees.

Core analysis in unconsolidated heavy oil sands is difficult because the core is soft and the oil changes as it cools and degasses. Laboratories usually measure fluid saturations by distillation-extraction (the Dean-Stark method), reporting the mass of oil, water and solids in each sample. Converting these to porosity and saturation requires the densities of the grains, oil and water, as the code in this lesson does. Net pay for a steamflood is then defined by cutoffs on oil content, porosity and clay volume from the logs, as in PET-350.

Clay layers, diatomite and injection limits

Three geological features decide much of a steamflood's performance. The first is clay layers. Thin, discontinuous clays let steam rise through the sand, which is useful because steam is buoyant. Continuous clays separate the reservoir into zones that must be steamed separately, and a steam zone that rises to a clay layer spreads along its base, overriding the oil below. The second is diatomite. In some west-side fields, such as Belridge and parts of Midway-Sunset, heavy oil sits in diatomite, a rock made of the silica skeletons of microscopic algae. Diatomite has very high porosity but very low permeability, so steam must be injected at high pressure to create fractures, which carries its own risks. The third is the injection pressure limit. Injecting steam above the pressure that fractures the rock and the overlying layers can create paths to the surface.

In 2019, a series of surface expressions, places where oil, water and steam flowed out of the ground, occurred at a field on the west side of the valley where cyclic steam was being injected at high pressure. CalGEM responded with stricter rules and a halt to new high-pressure cyclic steaming above the fracture gradient while it reviewed the practice. California's regulations for underground injection projects now require operators to demonstrate that injected fluids stay in the approved zone, and to monitor for surface expressions. The lesson for engineers is that the injection pressure limit is a geological limit as much as an equipment limit.

From core and logs to a development map

The development team turns the geology into numbers that a plan can use: maps of net pay, oil saturation, porosity, the depth to the base of the pay and the distance to any water or gas, and the location and continuity of shale or clay layers. These maps decide where well pairs or patterns go and how they are spaced. Delineation drilling on thermal leases is dense compared with conventional pools: an oil sands lease may have a vertical core hole every few hundred meters, and 3D seismic is used to map channel edges and mud plugs. On mature California leases, decades of well logs and temperature surveys provide the same kind of detail.

# From core: bitumen mass fraction to oil saturation, and net pay with a cutoff
SI = False  
rho_s, rho_o, rho_w = 2.65, 1.01, 1.00     # g/cm3 for quartz grains, oil and water
def saturations(wb, ww):
    # wb, ww: weight fractions of oil and water in the core sample; the rest is solids
    ws = 1.0 - wb - ww
    vs, vo, vw = ws / rho_s, wb / rho_o, ww / rho_w
    phi = (vo + vw) / (vs + vo + vw)
    return phi, vo / (vo + vw)
if SI: core = [(0.5, 0.135, 0.030), (1.5, 0.140, 0.028), (2.5, 0.120, 0.040), (3.5, 0.060, 0.090), (4.5, 0.115, 0.045), (5.5, 0.025, 0.150)]
else:  core = [(2.0, 0.135, 0.030), (5.0, 0.140, 0.028), (8.0, 0.120, 0.040), (11.0, 0.060, 0.090), (14.0, 0.115, 0.045), (17.0, 0.025, 0.150)]
step = 1.0 if SI else 3.0
cut = 0.06                     # 6 % bitumen by weight is a common net-pay cutoff for in situ projects
net = 0.0
unit = "m" if SI else "ft"
for depth, wb, ww in core:
    phi, so = saturations(wb, ww)
    pay = wb >= cut
    if pay: net += step
    print(f"{depth:5.1f} {unit} below top  oil {wb:5.1%} wt  porosity {phi:.2f}  So {so:.2f}  {'pay' if pay else 'not pay'}")
print(f"net pay in this {len(core) * step:.0f} {unit} interval: {net:.0f} {unit}")

  2.0 ft below top  oil 13.5% wt  porosity 0.34  So 0.82  pay
  5.0 ft below top  oil 14.0% wt  porosity 0.35  So 0.83  pay
  8.0 ft below top  oil 12.0% wt  porosity 0.33  So 0.75  pay
 11.0 ft below top  oil  6.0% wt  porosity 0.32  So 0.40  pay
 14.0 ft below top  oil 11.5% wt  porosity 0.33  So 0.72  pay
 17.0 ft below top  oil  2.5% wt  porosity 0.36  So 0.14  not pay
net pay in this 18 ft interval: 15 ft

Ethan's core interval shows a rich upper sand, a leaner silty layer and a water-wet sand at the base. The sample at 11 feet sits right at the cutoff. 'Lean layers like that let steam through but slow it down,' Grace says. 'It is the thick clay at the base that will steer the heat sideways.'

Seismic, delineation and the reservoir model

Core holes give exact information at points, but the space between them must be filled. In the Kern County fields, the density of wells is so high, often several per acre after a century of development, that log correlation between wells provides much of the information that seismic would give elsewhere. Geologists correlate sands and clay layers across thousands of wells, map net pay and oil saturation, and use temperature and saturation logs from observation wells to see where steam has already swept. They combine these into a 3D geological model, in which each cell carries a facies, porosity, permeability and oil saturation. That model is the input to thermal simulation and to the forecasts that support a development plan.

Uncertainty in the model is handled by building several realisations, each consistent with the data but different between wells. Running the forecast on several realisations shows how much the predicted rate or SOR could vary, and where an extra well would reduce that uncertainty most. That is the starting point of the risk register in lesson 12.

Testing the cap rock

Before steam is injected at higher rates, the strength of the confining layers and the stress in the ground must be known. Operators run step-rate tests, pumping fluid in increasing steps until the rock fractures and the pressure-rate curve changes slope, to find the fracture gradient. Core samples of the confining clays are tested for strength and permeability. Geomechanical models predict how the reservoir and overlying layers will expand as they heat and what injection pressure keeps a safety margin below fracture. In California, CalGEM's underground injection control rules require operators to demonstrate that injected fluids will stay in the approved zone and to keep injection pressures below the fracture gradient, and its rules on surface expressions set out the monitoring and response expected in steamed areas.

Containment work does not end with approval. Observation wells, surface heave monitoring and pressure surveillance continue for the life of the project, and any sign that steam or fluids are moving out of the target zone triggers investigation and, if necessary, lower injection pressures or a halt to steaming in the area.

Water, gas and other zones

The geologist also maps everything above and below the pay that could interact with steam. Water-bearing sands below the reservoir may supply bottom water that floods a chamber; water or gas zones above it may act as thief zones that drain heat and pressure; freshwater aquifers higher in the section must be protected. Old wells that penetrate the reservoir are another potential pathway, so their locations and condition are reviewed, and any that are not designed for thermal temperatures are repaired or replugged before steam reaches them. This inventory of zones and pathways is part of the application a regulator reviews before approving steam injection.

The result of all this work is a set of maps and models that the reservoir engineers, drilling engineers and regulators all rely on. The better they are, the better the wells are placed, the lower the steam-oil ratio and the more confident everyone can be that the steam will stay where it belongs.