WEBVTT

00:00:03.000 --> 00:00:13.000
Good morning everyone. Welcome today, number 3 of the northern California Earthquake Hazards  workshop. And we have our last in our series of sessions 

00:00:13.000 --> 00:00:17.000
focusing on a specific location in a multidisciplinary way.

00:00:17.000 --> 00:00:20.000
Today we are going to the Geysers and to lead us through this journey 

00:00:20.000 --> 00:00:26.000
 we have two people who are themselves Geysers of knowledge when it comes to geothermal areas.

00:00:26.000 --> 00:00:30.000
Denise Templeton and Ole Kaven. Woohoo![excitement]

00:00:30.000 --> 00:00:47.000
Good morning everybody as Sarah mentioned this morning session is focused on the Geysers, it's an exciting place and topic scientifically as earthquakes have long been welcomed in the geothermal  community as a tool to inform dynamics of the resource. The Geysers 

00:00:47.000 --> 00:01:04.000
specifically also interesting from a hazard perspective, because it's model of transparency and community engagement has been widely successful. Today's speakers will cover broad range of important topics and we're starting with the most important one or the most basic one maybe, where the heat's

00:01:04.000 --> 00:01:15.000
coming from. So, Jerry, take it away, please.

00:01:15.000 --> 00:01:23.000
I'll be talking about [indiscernible] Clear Lake in northern California.

00:01:23.000 --> 00:01:26.000
So I'll start out by talking a little bit about MT

00:01:26.000 --> 00:01:32.000
and get some regional context, then we will talk specifically about Clear Lake and I'll share some deep thoughts.

00:01:32.000 --> 00:01:39.000
So magnetotellurics is an electromagnetic method that measures the resistivity of the subsurface.

00:01:39.000 --> 00:01:45.000
Why is it important? Because electrical resistivity is directly sensitive to where fluids are have been in the crust.

00:01:45.000 --> 00:01:53.000
MT. Has a large dynamic range, meaning that we can measure orders of magnitude of frequency with one measurement.

00:01:53.000 --> 00:02:02.000
So that's usually say a 1,000 words to a 1,000 seconds, which gives you a depth range of tens of meters to tens of kilometers. Inherently,

00:02:02.000 --> 00:02:12.000
MT. has depth information as well as directional information because of the polarization of electromagnetic fields.

00:02:12.000 --> 00:02:21.000
MT is also a diffusive method, which means that energy is constantly being put into the ground, and we don't have to worry about Ray Pass.

00:02:21.000 --> 00:02:27.000
So again, we use Mt, because it's directly sensitive to where fluids are or have been in the crust.

00:02:27.000 --> 00:02:43.000
But Clear Lake, in a regional context, I'll show some preliminary results from resistivity model developed of the Western United States as part of various programs for the USGS.

00:02:43.000 --> 00:02:50.000
So here's the cross-section of that resistivity model on the left-hand side is west and to the right is east.

00:02:50.000 --> 00:02:55.000
We got an inset map showing where that cross-section goes, and it goes across Clear Lake.

00:02:55.000 --> 00:03:08.000
Clear lake here, on the left hand side, below this orange box we have the slab window, so the resistivity model here is showing blues is resistive, and red is conductive.

00:03:08.000 --> 00:03:17.000
So the slab window shows up as a resisted body, which indicates that there's not a large zone of partial melt below Clear Lake.

00:03:17.000 --> 00:03:31.000
That is part of the slab window, and if we compare that with seismic, we can see that the seismic shows a low velocity zone beneath Clear Lake that is quite distinct.

00:03:31.000 --> 00:03:38.000
But if we compare that with the resistivity model, we see that the resistivity model is sensitive to fluids, and the velocity is sensitive to temperature.

00:03:38.000 --> 00:03:49.000
So we're seeing that something is hot but it's above solidus that's at the solidest temperature that there isn't a bunch of partial melt below 

00:03:49.000 --> 00:03:55.000
Clear Lake, I mean there still can be, but on both it's an [indiscernible] temperature.

00:03:55.000 --> 00:03:58.000
So if we zoom in at Clear Lake we started clicking MT data up there about 6 years ago, starting at the Geysers.

00:03:58.000 --> 00:04:22.000
That project has expanded into a monitoring project where we're doing measurements over the course of 3 years to observe any spatial and temporal changes within the steam field. In conjunction with constant passive seismic tomography.

00:04:22.000 --> 00:04:26.000
As part of a California Volcano Observatory project 

00:04:26.000 --> 00:04:34.000
we are studying the clearly volcanic system, and as a whole and so we've collected a bunch of MT data up there.

00:04:34.000 --> 00:04:37.000
And today I'll show you preliminary results from a 3D 

00:04:37.000 --> 00:04:42.000
electrical resistivity model derives from these data.

00:04:42.000 --> 00:04:51.000
So first off, I'll just show some depth slices of the resistivity model again, blues are resistive, reds are conductive. On top,

00:04:51.000 --> 00:04:57.000
we have dots are earthquake hypocenters, colored by magnitudes.

00:04:57.000 --> 00:05:03.000
So dark colors are low magnitude, light colors are high magnitude, and for spatial reference 

00:05:03.000 --> 00:05:08.000
we have Clear Lake here in the middle we have some eruptive centers at these cones.

00:05:08.000 --> 00:05:13.000
So now MT. Helena, Cobb, MT. Hannah, MT. Konacti, Round mountain.

00:05:13.000 --> 00:05:17.000
The general trend is getting younger to the north east.

00:05:17.000 --> 00:05:23
The fault system is bounded by these two large faults, the Bartlett Springs fault and the Maacaama fault 

00:05:23.000 --> 00:05:32.000
 to the southwest. So this on the right is a reference map at what depth we're at.

00:05:32.000 --> 00:05:42.000
We'll just play this video and you can see, here the Geyser is quite seismically active.

00:05:42.000 --> 00:05:51.000
But we do have the decent amount of earthquakes along the faults. So below about 5 kilometers, you'll see there's not as much seismicity 

00:05:51.000 --> 00:06:00.000
the Geysers, and as you get down to about 10 kilometers, you start to see there was seismicity following these faults. Here's the Bartlett Springs fault.

00:06:00.000 --> 00:06:08.000
Then there's a gap of seismic activity, and then we start to see seismic activity again 

00:06:08.000 --> 00:06:21.000
along the Berryessa. So this conductive anomaly here somehow interrupting seismic activity, suggesting that it's probably hot as some fluids in it, and likely partial melts.

00:06:21.000 --> 00:06:27.000
And as we go deeper you'll start to see around 15 to 20 kilometers.

00:06:27.000 --> 00:06:30.000
You start to see these clusters on the northwestern side in the south.

00:06:30.000 --> 00:06:38.000
Sorry the northeastern side and the southwestern side of Clear Lake and these have been identified as long period events.

00:06:38.000 --> 00:06:45.000
And it's common that they occur on the peripheral volcanic systems and conductive anomalies.

00:06:45.000 --> 00:06:54.000
You often see that same in the Long Valley.  

00:06:54.000 --> 00:07:05.000
So if we take a closer look at one of the cross-sections; here's a cross-section looking south to north on the left- hand side is south in the right-hand side is north. We've got some eruptive centers up here at the top.

00:07:05.000 --> 00:07:12.000
Designated by these cones and the general trend, again, is from older to younger 

00:07:12.000 --> 00:07:20.000
 as you move north. If we put some faults on here, you can start to see that the conductive anomalies are controlled by the faults.

00:07:20.000 --> 00:07:25.000
So we have the Sulfur Creek fault and the Konacti fault bounding

00:07:25.000 --> 00:07:28.000
the Geyser's Plutonic Complex, which is this 

00:07:28.000 --> 00:07:34.000
 resisted club. Underneath Cobb Mountain, we have a conductive body under Mt. Hannah 

00:07:34.000 --> 00:07:38.000
 that's fault controlled. Konacti seems to follow it.

00:07:38.000 --> 00:07:48.000
A fault as well, and then we have the Bartlett Springs out to the far north, and if we put earthquake hypocenters on top, you can see the Geysers Plutonic

00:07:48.000 --> 00:07:56.000
Complex is quite active. Sulfur Creek fault here is seismically active.

00:07:56.000 --> 00:07:57.000
We have seismicity within this conductive region which suggests that it's not partial melt 

00:07:57.000 --> 00:08:03.000
that it's probably aqueous fluids, or some other conducting phase 

00:08:03.000 --> 00:08:15.000
 that's causing that conductive anomaly, and under Konacti you see a distinct bolt that could have been the pathway for magma to ascend to the surface.

00:08:15.000 --> 00:08:21.000
Then to the north, you see this cluster of long period events.

00:08:21.000 --> 00:08:27.000
If we look at a different cross section, this one south, west to northeast again, we have some eruptive centers.

00:08:27.000 --> 00:08:35.000
Now we have Round Mountain, which is a cinder cone, and put on some faults.

00:08:35.000 --> 00:08:41.000
You can see that control of those faults on the conductivity structures.

00:08:41.000 --> 00:08:49.000
Again, we have the Geysers here. Geysers Plutonic Complex controlled by the Silver Creek from the Koliami fault.

00:08:49.000 --> 00:09:03.000
Mt. Hannah seems to be quite controlled by these two faults, as well as Round Mountain seems to be controlled by Bartlett Springs fault and a related fault, and if we put on seismicity, you can see the Geysers lights up quite a bit. Again,

00:09:03.000 --> 00:09:09.000
we have, seismicity within this conductive region, suggesting that it's not partial melt.

00:09:09.000 --> 00:09:12.000
We have some seismicity on Round Mountain, that's quite interesting.

00:09:12.000 --> 00:09:30.000
It follows the boundary between productive anomaly and resistive anomaly. And then another feature to observe is that there's seismicity within this large resistive zone underneath the Geyser suggesting that again, that it is not a bunch of partial melt.

00:09:30.000 --> 00:09:40.000
How does this compare to seismic, the seismic tomography model came out last year, so on the left we have Vp, Vs resistivity, and then density.

00:09:40.000 --> 00:09:48.000
And then each row designates a different depth slice of 4 kilometers, 8 kilometers in the middle of 16 kilometers at the bottom.

00:09:48.000 --> 00:09:55.000
We have the outline of Clear Lake, and the red circle designates where the low velocity zones occur.

00:09:55.000 --> 00:10:15.000
So 4 kilometers, you see similar patterns in all four low resistivity, low density, low velocity. Suggests that maybe there is a bit of partial melt, could be sediments, could be temperature. At 8 kilometers, it's 

00:10:15.000 --> 00:10:18.000
 a bit more distinct in the velocity, in the resistivity 

00:10:18.000 --> 00:10:30.000
 it's a little bit faded, suggesting that there's not as much partial melt as, say, the density or the velocity anomaly would suggest and it's more or less gone at 16 kilometers.

00:10:30.000 --> 00:10:34.000
But there is this conductive anomaly to the north 

00:10:34.000 --> 00:10:38.000
 that splits the seismicity of the Bartlett Springs, and various of faults which it suggests that it's probably hot  and includes fluids 

00:10:38.000 --> 00:11:04.000
Could be partial melt. And if we look at a different cross-section now similar to what you had before for the resistivity, so on the top we have the velocity P-wave and S-wave, and then the resistivity and density, the Geysers, here is a low [indiscernible] a high velocity

00:11:04.000 --> 00:11:09.000
 Zone high BSF, and a high resistivity zone.

00:11:09.000 --> 00:11:16.000
But then to the north east, there's this low velocity zone that is highlighted in the P.

00:11:16.000 --> 00:11:22.000
 and S also comes up as a low or high conductivity 

00:11:22.000 --> 00:11:25.000
 anomaly, and a low density anomaly.

00:11:25.000 --> 00:11:29.000
So this is suggesting that there could be partial melt here.

00:11:29.000 --> 00:11:39.000
But not as much as say, density alone, or seismic velocity alone would suggest.

00:11:39.000 --> 00:11:51.000
The preliminary results is pretty consistent between the geophysical models that the eruptive centers of Clear Lake are fault controlled.

00:11:51.000 --> 00:11:59.000
From the resistivity model, it seems evident that there's a larger body beneath the Geysers 

00:11:59.000 --> 00:12:04.000
Plutonic Complex, could be a larger plutonic body.

00:12:04.000 --> 00:12:05.000
It's quite hot, but right at the [indiscernible]

00:12:05.000 --> 00:12:11.000
- at solid temperatures. So it's could have personal melt,

00:12:11.000 --> 00:12:18.000
 but on bulk it's not hot enough to be melted.

00:12:18.000 --> 00:12:30.000
And then there could be a possible magmatic body to the northeast, between the Bartlett Springs and the Berryessa  fault. So that brings up some other questions.

00:12:30.000 --> 00:12:31.000
Is the slab window ahead of the eruptive centers?

00:12:31.000 --> 00:12:46.000
It seems like there's a zone of partial melt to the northeast if it's related to the volcanic system, how does that magma transport to the surface?

00:12:46.000 --> 00:12:53.000
Or are there other sources of magma or mechanisms to get magma to the surface?

00:12:53.000 --> 00:13:04.000
It seems that eruptive centers are quite fault controlled, so is there a causal relationship between fault ruptures and magmatic activity?

00:13:04.000 --> 00:13:08.000
Is it that fault ruptures and magma ascends?

00:13:08.000 --> 00:13:12.000
Or is it magma ascending causing fault ruptures.

00:13:12.000 --> 00:13:18.000
And then the final question is why are the Geysers where they are?

00:13:18.000 --> 00:13:26.000
It seems that it's quite serendipitous that that plutonic body found a weakness in the crust.

00:13:26.000 --> 00:13:30.000
To ascend the close enough to the surface to develop a steam field.

00:13:30.000 --> 00:13:35.000
And why didn't it happen anywhere else in the Clear Lake Volcanic System?

00:13:35.000 --> 00:13:36.000
That's all I had so I'll be happy to take any questions.

00:13:36.000 --> 00:13:41.000
Thank you.

00:13:41.000 --> 00:13:52.000
Great thanks, Jared. Thanks for starting us out next up is Patricia Martinez Garzón who's going to talk about the induced earthquake potential and geothermal resources.

00:13:52.000 --> 00:14:01.000
Hi there! Thank you for watching this presentation online. I am Patricia Martinez Garzón, Seismologist from GFZ-Potsdam in Germany.

00:14:01.000 --> 00:14:11.000
I would like to present to you our study, entitled, "Induced Earthquake Potential in Geothermal Reservoirs, Insights from the Geysers in California."

00:14:11.000 --> 00:14:20.000
I'm starting by acknowledging the co-authors of this study, Grzegorz Kwiatek, Stephan Bentz, Marco Bohnhoff, and Georg Dresen.

00:14:20.000 --> 00:14:30.000
All of us from GFZ-Potsdam section 4.2 geomechanics and scientific drilling.

00:14:30.000 --> 00:14:34.000
So let's start with the motivation for this talk.

00:14:34.000 --> 00:14:52.000
The thermal reservoir production and associated induced seismicity may experience a pronounced attention in a few years, given the ambitious plans for reducing greenhouse gas emissions towards a carbon neutral economy and society. Very common by product of these activities is the

00:14:52.000 --> 00:15:04.000
occurrence of induced seismicity as a green energy source geothermal operations generally face little public acceptance issues, but this can instantly change when induced 

00:15:04.000 --> 00:15:10.000
 seismic events are large enough to be felt by the local population or cost damage to infrastructure.

00:15:10.000 --> 00:15:30.000
In this study our main goals are first of all to summarize some of the most important lessons learned from the analysis of micro-seismicity at the Geyser Geothermal Field, from different studies that we have performed over the last few years; and secondly, to start identifying key seismological

00:15:30.000 --> 00:15:43.000
 features, such as the seismic injection efficiency, that, when combined with injection data may enable a better quantification of the seismic hazards.

00:15:43.000 --> 00:15:55.000
I assume that many of you are already familiar with the Geysers Geothermal Field in northern California, which is the largest production reservoir worldwide.

00:15:55.000 --> 00:16:06.000
The geothermal power at the Geysers started around the 1960s by extracting steam from the reservoir, which is approximately located between 2 and 4 kilometers.

00:16:06.000 --> 00:16:12.000
 depth. Currently more than 20,000 seismic events are recorded 

00:16:12.000 --> 00:16:18.000
 annually, which are here on these orange dots, with magnitudes up to about 4.9. We focused 

00:16:18.000 --> 00:16:20.000
 our analysis on the isolated cluster of seismicity in the northwestern part of the field.

00:16:20.000 --> 00:16:30.000
Here you see the map view and this is the latitude versus North- 

00:16:30.000 --> 00:16:36.000
West step, and in this part of the field only two wells are injecting fluid 

00:16:36.000 --> 00:16:51.000
 in the nearby 39, and 29, with 2 well traces, and the relocated catalogue from this region contains about 1,600 seismic events.

00:16:51.000 --> 00:17:00.000
So we applied a number of processing techniques to increase the data resolution and to obtain additional seismo- mechanical parameters.

00:17:00.000 --> 00:17:12.000
The most important ones that we use are the refinement of earthquake locations for imaging the fluid migration pads and potential pre-existing seismogenic faults at depth.

00:17:12.000 --> 00:17:31.000
The estimation of seismic moment tensors to unravel the fracture network and their respective contribution of sharing crack opening or crack closing included in the earthquake source. The earthquake source parameters to estimate the strength of individual

00:17:31.000 --> 00:17:36.000
earthquakes in terms of their [indiscernible] moment and stress drop.

00:17:36.000 --> 00:17:44.000
Finally, they intervene at times and distances between the earthquakes to identify background and triggered seismicity.

00:17:44.000 --> 00:17:50.000
That means foreshocks and aftershocks.

00:17:50.000 --> 00:18:09.000
And during the entire analyzed time period of about 8 years, the seismicity rates followed closely the combined injection rates of the wells P-9 and P-29, are here represented with different colors in contrast to the production from the five

00:18:09.000 --> 00:18:14.000
nearest producing wells showed no significant link to seismicity.

00:18:14.000 --> 00:18:30.000
This seismicity occurred distributed in an ellipsoidal volume containing the injection wells, and the special extent of the seismicity from the well P-29 varied seasonally according to the volume of water injected. With larger volumes of water injected

00:18:30.000 --> 00:18:35.000
 in winter, resulting in a larger activated volume display in seismicity.

00:18:35.000 --> 00:18:40.000
Here you see the seismicity from the analyzed cluster 

00:18:40.000 --> 00:18:50.000
 in map view, color encoded with that type of faulting style and rather than the activation of one single fault orientation 

00:18:50.000 --> 00:18:56.000
 the larger diversity obtained in the focal mechanisms indicated the activation of our network of

00:18:56.000 --> 00:19:12.000
 fractures and faults. The epicenters of the seismicity were mostly contained in an ellipse with its long axis almost aligned with the orientation of local maximum horizontal stress estimated from version of focal

00:19:12.000 --> 00:19:18.000
mechanisms. These findings are very consistent with other EGS 

00:19:18.000 --> 00:19:26.000
site projects such as sulphurate in France.

00:19:26.000 --> 00:19:31.000
The maximum observed magnitudes within the analyzed part of the Geyser

00:19:31.000 --> 00:19:40.000
 Geothermal is 3.2. We found that in this area the largest events tend to locate near the open hole section of the injection 

00:19:40.000 --> 00:19:53.000
well, P-29, where stress perturbations from the fluid injection are expected to be the largest. A semi-empirical relationship between the cumulative injected volume and the maximum observed 

00:19:53.000 --> 00:20:02.000
 seismic moment release was found by Art McGarr to provide an upper bound for a number of fluid injection operations. For the fluid volume 

00:20:02.000 --> 00:20:05.000
 injected into the reservoir from the beginning until approximately 2.5 years after 

00:20:05.000 --> 00:20:29.000
 where the largest event occurred McGarr's model will predict a maximum observed magnitude of 5, which is larger than what has ever been observed at the Geysers, and therefore, they observe the value of 3.2 is significantly smaller compared to the predicted value and McGarr's model

00:20:29.000 --> 00:20:48.000
seems to not apply very well. This is not entirely surprising, as the McGarr's model applies to seismic events that are caused by fluid pressure perturbation, but at the Geysers, the cold water injected into the high temperature rock resulting in elevated thermal stresses causing

00:20:48.000 --> 00:20:56.000
 seismicity, and as substantial proportion of the injected water is rapidly converted to steam.

00:20:56.000 --> 00:21:15.000
In addition, the low observed the maximum magnitude could also be explained by a prominent role that the thermal stresses play at the Geysers, and the possibility of additional aseismic deformation related to the fluid injection activities.

00:21:15.000 --> 00:21:33.000
In contrast, we found a very good agreement between the observed maximum magnitude from our project 3.2, and its prediction by the relation proposed by Shapiro, based on the extent of the perturbed volume from the injection as you can see here. Which in this

00:21:33.000 --> 00:21:37.000
case predicted a maximum magnitude of 3.4.

00:21:37.000 --> 00:21:48.000
In our case we found an excellent correlation that you see here between the injection rates and the volume in which seismicity occurs.

00:21:48.000 --> 00:22:09.000
Therefore, this indicated that in this area of the Geysers thermal field the maximum magnitude appears to be better constrained by the size of the seismicity cloud, which in terms relate to the injection rate not to the injection volume.

00:22:09.000 --> 00:22:14.000
So earthquake static stress drop represents a seismically derived proxy for the shear stress 

00:22:14.000 --> 00:22:17.000
released during an earthquake normalized by its fault area. At the reservoir scale 

00:22:17.000 --> 00:22:42.000
 A statistically significant inverse dependence between calculated stress drops and injection rates was found first for the project in Vassel, for susefared and also for [indiscernible] similar results obtained at the Geysers and 

00:22:42.000 --> 00:23:06.000
 varying stress drop being inversely proportional to pore pressure has been suggested to indicate unstable versus stable rupture of the same asperity. Therefore, pointing towards joined seismic and aseismic slip release.

00:23:06.000 --> 00:23:16.000
Induced earthquakes related to enhanced the geothermal systems commonly include significant positive non-[indiscernible] components, indicating the opening of new fractures.

00:23:16.000 --> 00:23:32.000
And this result in permeability enhancement, we sorted our seismicity temporarily, and calculated the average isotopic percentage representing volumetric changes at the earthquake source for moving windows containing 20 events.

00:23:32.000 --> 00:23:51.000
We found that the aggregate in the injection rates from the two wells, which is here represented in red correlates very well with the average positive, isotropic momentizer components of the induced earthquakes, which is here represented in red. The 

00:23:51.000 --> 00:23:55.000
components were larger than the estimated uncertainties.

00:23:55.000 --> 00:24:08.000
We concluded that the increased isotropic components during peak injections call represents the opening of new fractures during time periods of increased poor fluid pressure. On the right,

00:24:08.000 --> 00:24:16.000
We have a figure showing the depth versus time distribution of the isotropic components, which is here color encoded.

00:24:16.000 --> 00:24:30.000
We observe a progressive temporal increase in the average amount of isotropic component which was in interpreted to be a proxy for the slow reservoir permeability increase in this portion of the field.

00:24:30.000 --> 00:24:39.000
The largest volumetric components do not concentrate at the borders of the stimulated volume, but rather within the main bulk of the cloud.

00:24:39.000 --> 00:24:49.000
At around 400 meters from the open hole section of the well P-29.

00:24:49.000 --> 00:24:57.000
The v-value from the Gutenberg Richard distribution is a very important parameter for seismic hazard models.

00:24:57.000 --> 00:25:14.000
The analysis of the long-term temporal trend of the b-value in the northwestern part of the Geysers revealed that the b-value remained approximately constant in space and time at about 1.22 with small short-term fluctuations.

00:25:14.000 --> 00:25:23.000
Therefore, we concluded that the seismic hazard here is rather controlled by the seismic activity rate rather than by changes in the magnitude-

00:25:23.000 --> 00:25:24.000
 frequency distribution. We also applied the nearest neighbor approach to calculate [indiscernible] times and distances of the space 

00:25:24.000 --> 00:25:38.000
 seismicity. The analysis revealed that most of the seismicity is represented by background events, and we obtained only a minor proportion of trigger 

00:25:38.000 --> 00:25:51.000
 seismicity which is all here represented with the red and blue bars, and however, we did obtain a larger proportion of foreshocks with respect to aftershocks.

00:25:51.000 --> 00:26:03.000
 during peak injection time periods.

00:26:03.000 --> 00:26:08.000
Let's talk a little about the energy balance of geothermal projects.

00:26:08.000 --> 00:26:18.000
Assuming that the tectonic stresses are negligible, the input energy sources are the hydraulic energy defined as the product of the injected volume 

00:26:18.000 --> 00:26:24.000
 multiplied by the average reservoir per pressure, and, second, the thermal energy created by the injection of cold water in the hot reservoir rock.

00:26:24.000 --> 00:26:44.000
These input energies are consumed into radiated seismic energy invested in the aseismic deformation and the energy required to hit the reservoir fluids and produce steam.

00:26:44.000 --> 00:26:58.000
In this context, I will like to introduce the seismic injection efficiency, which is defined as the ratio between the hydraulic energy of the system and the energy that is radiated through seismic waves.

00:26:58.000 --> 00:27:03.000
This slide we show the evolution of the seismic injection efficiency.

00:27:03.000 --> 00:27:23.000
During the stimulation of several geothermal projects which are listed here with respect to the hydraulic energy of the systems, and also with respect to the cumulative volume injected. In red, we have the temporal evolution of the seismic injection efficiency at

00:27:23.000 --> 00:27:32.000
The Geyser is thermal field. The first thing that we notice is that this parameter tends to remain approximately constant with time.

00:27:32.000 --> 00:27:38.000
If the system behaves in a stable manner and the earthquakes are self-arrested.

00:27:38.000 --> 00:27:39.000
This was her first time noticed in the Helsinki St-1

00:27:39.000 --> 00:27:54.000
Deep hit project, which is here. The green line. However, some projects show a deviation from this stable system.

00:27:54.000 --> 00:28:15.000
This is detected by an increase in the seismic injection efficiency with time, such as the black line here representing the copper basin in Australia. The purple line here representing the Pohang product that shows a rapid increase.

00:28:15.000 --> 00:28:33.000
These observations led us to conclude that monitoring of near real-time seismicity and seismic injection efficiency could help identifying your thermal reservoirs that are more prone to large earthquakes or to unstable runaway raptor.

00:28:33.000 --> 00:28:53.000
However, it is also important to emphasize that the monitoring of the seismic injection] efficiency does not imply that all unstable ruptures may not occur, such as due to the transfer of stresses between earthquakes or due to other factors that can

00:28:53.000 --> 00:29:01.000
modify the field, and then move the system towards a raptor.

00:29:01.000 --> 00:29:12.000
So this is basically all in summary, we describe the key results from the analysis of micro-seismicity and geomechanical parameters that Geyser Geothermal Field.

00:29:12.000 --> 00:29:30.000
We mainly found that monitoring temporal evolution of seismic injection efficiency may allow to identify a general propensity of the reservoir to nucleate unstable earthquakes of increased seismic hazards and the analysis at the Geysers when compared to other

00:29:30.000 --> 00:29:52.000
Geothermal sites allows us to hypothesize that certain features identified in the seismicity may point towards lower and high potential interruptor in unstable or stable earthquakes.  I would like to leave this table here for people to compare it we looked at different parameters such

00:29:52.000 --> 00:29:53.000
As the b-value. Where does the deformation 

00:29:53.000 --> 00:30:16.000
occurs whether the seismic injection efficiency follows stationary trend, or it increases with time, whether there are no double couples components or not, and the level of earthquake interaction. So

00:30:16.000 --> 00:30:28.000
So this was mainly all just to say that this work was published in the special issue of Geothermal by the Living Edge in 2020, and I leave it here with some references.

00:30:28.000 --> 00:30:48.000
I would like to take this opportunity to thank the Northern California Earthquake Data Center for their huge support and feedback while working with the Geysers data over these last years, and I will be happy to reply to any questions that you have or to engage in the discussion on this topic. Thank you.

00:30:48.000 --> 00:30:58.000
Thanks, Patricia. Next up is Andrew Delorey, who's gonna talk about triggered earthquakes as a probe for stress.

00:30:58.000 --> 00:31:12.000
Hello! My name is Andrew Delorey from Los Alamos National Laboratory and today I'm going to be talking about using triggered earthquakes as a probe for stress conditions on faults.

00:31:12.000 --> 00:31:15.000
I'm going to be talking about two study areas today.

00:31:15.000 --> 00:31:20.000
The first is the Geyser's Geothermal Field in California, just north of San Francisco.

00:31:20.000 --> 00:31:30.000
This is the largest Geothermal Field in the United States in terms of power generation. Powers generated by extracting steam from the subsurface.

00:31:30.000 --> 00:31:34.000
This Geothermal Field has been generating power since 1960.

00:31:34.000 --> 00:31:44.000
Initially the field was replenished by reinjecting steam condensate, but this became insufficient for replenishing the system.

00:31:44.000 --> 00:31:50.000
So plant operators began injecting treated wastewater to maintain steam production.

00:31:50.000 --> 00:31:57.000
The water table falls below the depth of most of the earthquakes, so this can be thought of as a water poor system.

00:31:57.000 --> 00:32:03.000
The black dots on the map to the right indicate earthquakes from one of the catalogs that I will be discussing today.

00:32:03.000 --> 00:32:09.000
The cold water injections result in thermal elastic stresses that trigger earthquakes. During cold water 

00:32:09.000 --> 00:32:19.000
 injections, the faulting type transitions from normal to strike-slip as a result of the thermal contraction in the rock.

00:32:19.000 --> 00:32:36.000
The second study area is in Oklahoma. Most seismicity in Oklahoma is induced due to wastewater injections associated with oil and gas operations. Observations show that injection rates volumes proximity to wells and preexisting

00:32:36.000 --> 00:32:47.000
structures all contribute to seismicity. Wastewater is typically injected into the Arbuckle Formation and then some of it migrates into the basement below.

00:32:47.000 --> 00:32:59.000
This can be thought of as a water-rich system, as pore fluid pressure has a strong impact on seismicity. Faulting is mainly strike-slip with some normal faulting in the North.

00:32:59.000 --> 00:33:03.000
Here are the earthquake catalogs I will be discussing today.

00:33:03.000 --> 00:33:16.000
There are two catalogs for the Geysers and one for Oklahoma. Since I'll be projecting solid earth tidal stresses to these earthquakes, I'll be using the actual fault plane solution for the Berkeley Mechanism Catalog and for the other two

00:33:16.000 --> 00:33:21.000
catalogs, I will assume that the earthquakes are optimally oriented in the regional stress field.

00:33:21.000 --> 00:33:27.000
This is consistent with observations and more Coulomb faulting theory.

00:33:27.000 --> 00:33:34.000
When title stresses are projected onto a fault plane there is a fault normal component in a fault shear component.

00:33:34.000 --> 00:33:42.000
This figure shows title stresses projected onto the fault plane solution for the Prague earthquake in Oklahoma.

00:33:42.000 --> 00:33:47.000
The blue line indicates the fault normal stress and the red line indicates the fault 

00:33:47.000 --> 00:33:54.000
 shear stress. Two important features of these curves or that the fault normal stress has a much greater magnitude 

00:33:54.000 --> 00:34:12.000
 then the fault shear stress in that the two curves are at a phase by about 90 degrees. When modeling Coulomb stress we use one of the two equations shown here for the anisotropic and isotropic poor elastic conditions. Most faults are likely to have anisotropic pore elastic

00:34:12.000 --> 00:34:29.000
 conditions due to the fabric within the sheer zone of the fault. We can see that both fault normal and fault shear stress contribute to Coulomb stress, but that the contribution of fault normal stress depends on the pore pressure response and the coupling between pore

00:34:29.000 --> 00:34:43.000
 pressure and confining stress. For the B term which is Skempton's coefficient a value near one indicates a strong coupling between pore pressure and confining stress, which results in near 0

00:34:43.000 --> 00:34:49.000
 contribution of the fault normal stress to the change in Coulomb stress.

00:34:49.000 --> 00:34:54.000
The B-value near 0 means low coupling between pore pressure and confining stress, which results in the fault 

00:34:54.000 --> 00:35:15.000
normal stress dominating the change in Coulomb stress by virtue of its much higher amplitude. Since the fault normal and fault shear stress are at a phase, we can test the correlation between fault normal stress and fault shear stress with seismicity separately.

00:35:15.000 --> 00:35:18.000
We would like to quantify earthquake triggering.

00:35:18.000 --> 00:35:25.000
In this case we are testing if there's a relationship between earthquake occurrence and the stress magnitude. I'm going to introduce a Modified Quantile-Quantile plot or MQQ.

00:35:25.000 --> 00:35:35.000
I use two different normalized cumulative distribution functions.

00:35:35.000 --> 00:35:45.000
The first one relates time versus stress. This means that the distribution represents how much time is spent at each stress level or less.

00:35:45.000 --> 00:35:56.000
The second one relates earthquakes versus stress. This means that the distribution represents how many earthquakes occur at each stress level, or less.

00:35:56.000 --> 00:36:17.000
If there is no relationship between earthquakes and stress, these two normalized cumulative distribution functions plotted against each other should produce a straight line between 0,0 and 1,1 that is the number of earthquakes that occur at each stress level or less should be equal to the amount of time spent at each stress

00:36:17.000 --> 00:36:29.000
 level or less. If the line falls below this diagonal, it means that there are fewer earthquakes than expected at a given stress level or less.

00:36:29.000 --> 00:36:34.000
This indicates that there are more earthquakes than expected at higher stress levels 

00:36:34.000 --> 00:36:43.000
 then if there was no relationship between earthquake occurrence and stress. In the upper right, I have shown the QQ

00:36:43.000 --> 00:36:48.000
 plot for 100 realizations of a Poisson catalog.

00:36:48.000 --> 00:36:53.000
You can see that it looks like a straight line between 0,0 and 1,1 if

00:36:53.000 --> 00:36:58.000
 we remove the slope and vertically exaggerate the axes 

00:36:58.000 --> 00:37:02.000
 we can see that there is some random variability in the 100 realizations.

00:37:02.000 --> 00:37:08.000
 but that their average shown in yellow is very close to a horizontal line.

00:37:08.000 --> 00:37:24.000
 as expected. If a curve falls outside of these 100 realizations we could reject a null hypotheses that there is no relationship between earthquake occurrence and stress at a confidence level greater than 99%.

00:37:24.000 --> 00:37:28.000
Here are some examples of non-homogeneous Poisson catalogs.

00:37:28.000 --> 00:37:29.000
The Poisson rate is no longer constant, but is the sum of a constant rate,

00:37:29.000 --> 00:37:44.000
 plus a variable rate, that is represented by the tidal stress function. On the left is shown an earthquake catalog with 20% forcing.

00:37:44.000 --> 00:37:49.000
This means that the magnitude of the variable rate is 20% of the long-term average rate.

00:37:49.000 --> 00:37:56.000
The middle is shown a catalog with 30% forcing, on the right is an earthquake catalog with 30% forcing, 

00:37:56.000 --> 00:38:01.000
 but the earthquake rate is related to the stress squared instead of just stress.

00:38:01.000 --> 00:38:09.000
This skews the curve to the right. This means that increasing stress becomes increasingly effective at triggering earthquakes.

00:38:09.000 --> 00:38:10.000
We could devise another relationship where increasing stress becomes decreasingly 

00:38:10.000 --> 00:38:22.000
 effective at triggering earthquakes which would skew the curve to the left.

00:38:22.000 --> 00:38:28.000
Since the real catalogs are not Poisson catalogs, because they have aftershocks 

00:38:28.000 --> 00:38:34.000
 we need to compare them to realizations of catalogs that have the same characteristics as the real catalogs.

00:38:34.000 --> 00:38:50.000
We do this by creating permutations of the real catalogs. A permutation is constructed by identifying the largest inter-vent times in the catalogs, and then segmenting the catalogs by the longest center event times using a number of segments equal to

00:38:50.000 --> 00:39:08.000
1% of the number of earthquakes. Using many fewer segments limits the number of firm mutations we can produce, and using many more segments increasingly destroys the characteristics of the original catalog. After segmentation, we randomly reorder the segments.

00:39:08.000 --> 00:39:13.000
Making the catalog circular in time, then select a new random starting point.

00:39:13.000 --> 00:39:21.000
These permutations preserve most of the characteristics of the original catalog, and aftershock sequences are preserved.

00:39:21.000 --> 00:39:29.000
Some long-term dependencies may be destroyed, but these would be months or longer.

00:39:29.000 --> 00:39:33.000
So why am I not using previous statistical methods? At the top 

00:39:33.000 --> 00:39:42.000
I'm showing the results for a one-year long Poisson catalog with an earthquake rate of one per day in a title forcing a 50%.

00:39:42.000 --> 00:39:46.000
We can see on the left that the Schuster Test does not detect this tidal

00:39:46.000 --> 00:40:05.000
 forcing, while on the right the MQQ Test does detect the tidal forcing because the actual catalog falls outside of 100 realizations of a homogeneous Poisson catalog. The Schuster Tests determined significance at each period independently, while the MQQ Test considers correlation

00:40:05.000 --> 00:40:22.000
to the entire tidal function. This is why it is more sensitive than the Schuster Test. The Schuster Test starts to detect tidal forcing only for much bigger catalogs or much stronger forcing. There are many previous studies both positive and negative that use the Schuster Test for tidal

00:40:22.000 --> 00:40:25.000
stresses, but this test is poorly suited for the task.

00:40:25.000 --> 00:40:31.000
Do not use the Schuster Test for tidal forcing. Other tests approximate 

00:40:31.000 --> 00:40:37.000
 tidal function as a simple sine function, or rely on subjective binning of earthquakes.

00:40:37.000 --> 00:40:46.000
These are poor approximations. Only the MQQ Test uses the model tidal function without approximations.

00:40:46.000 --> 00:40:51.000
So let's look at the results. For the three catalogs I introduced earlier before I get into the figures.

00:40:51.000 --> 00:40:52.000
I just want to note that for the Geysers catalogs, I'm presenting the results for fault 

00:40:52.000 --> 00:40:57.000
 normal stress, and for the Oklahoma catalog, I'm presenting the results for the fault 

00:40:57.000 --> 00:41:04.000
 Shear stress. I will show the fault results later in a table. The Poisson catalog shown here is an explicitly referring to fault normal or fault 

00:41:04.000 --> 00:41:15.000
 shear stress. Just the tidal stress function with 20% forcing, it could be either kind of stress.

00:41:15.000 --> 00:41:21.000
The blue curves are the results for 100 permutations of the respective catalogs.

00:41:21.000 --> 00:41:36.000
The yellow curve is the average of the permutations and the red curves are the actual catalogs. We can see that permutations of catalogs with aftershocks, such as the three rail catalogs used here and permutations of catalogs without aftershocks such as

00:41:36.000 --> 00:41:40.000
the Poisson catalog do not exhibit tidal forcing.

00:41:40.000 --> 00:41:55.000
This is because, aftershocks do not masquerade as tidal forcing. We can see here that the EGSC catalog is highly correlated with tidal stress, with a confidence level much greater than 99%. The BMC in Oklahoma catalogs

00:41:55.000 --> 00:42:00.000
are also correlated with tidal stress with the confidence of around 99%.

00:42:00.000 --> 00:42:07.000
Since the red curves are similar to the biggest sitelire of the 100 permutations.

00:42:07.000 --> 00:42:08.000
We can estimate the level of forcing in the 

00:42:08.000 --> 00:42:19.000
EGSC in Oklahoma catalogs by comparing them to realizations of Poisson catalogs with tidal forcing. The Poisson catalogs we are using for comparison 

00:42:19.000 --> 00:42:34.000
 don't have aftershocks like the real catalogs, but we saw the aftershocks don't masquerade as tidal forcing. For the Poisson catalogs with forcing we match just by eye the mean of 100 realizations shown in yellow to

00:42:34.000 --> 00:42:40.000
 the real catalogue in red, based on different levels  of forcing. For all EGSC earthquakes.

00:42:40.000 --> 00:43:02.000
 we estimate an 18% level of forcing. We can look at only the deepest 50% of the earthquakes where the best match is 17% forcing in only the 50% shallowest earthquakes where the best match is 27%. For Oklahoma the best match is 60%.

00:43:02.000 --> 00:43:06.000
We also apply to similar test where we just compare the likelihood values for the actual catalogs versus permutations of the actual catalogs encountered 

00:43:06.000 --> 00:43:29.000
 how frequently the permutations of a higher likelihood value then the actual catalog. To calculate the likelihood value, we simply assign a stress value to each earthquake based on when it occurs as shown in the figure here for fault normal stress. Then, we add up all

00:43:29.000 --> 00:43:41.000
the values for the entire catalog and its permutations, we can determine how often permutations of the catalog have a higher likelihood score than the real catalog.

00:43:41.000 --> 00:43:49.000
Here are the results. At the top of this table are the headings for the study areas in earthquake catalogs.

00:43:49.000 --> 00:43:55.000
Look for the row that begins percentage of permutations that are more correlated than the real catalog.

00:43:55.000 --> 00:44:11.000
If we look across the sub-row that says title share stress, we can see that 75% of the permutations of the EGSC catalog have a higher likelihood score than the real catalog. The number is 67% for the BMC catalog but only 

00:44:11.000 --> 00:44:17.000
0.05% for the Oklahoma catalog. For the several labeled fault normal stress 

00:44:17.000 --> 00:44:19.000
 we can see that 0% of the permutations of the EGSC catalog had a higher likelihood score than the real catalog.

00:44:19.000 --> 00:44:39.000
This was for 2,000 permutations. The number is 0.3% for the BMC catalog in 63% for the Oklahoma catalog. We can subdivide the BMC catalog by depth and plunge since we have fault plane solutions for this

00:44:39.000 --> 00:44:40.000
 catalog. For the deepest 50% of earthquakes;

00:44:40.000 --> 00:44:59.000
7% of permutations have a higher likelihood score, but less than 1% for the shallowest 50% of the earthquakes. For the lowest 50% of plunges, 13% of permutations had a higher likelihood score than the real catalog, but less than 1% of

00:44:59.000 --> 00:45:04.000
the 50% highest plunges. So for the BMC catalog shallow normal faulting events are most highly sensitive to tidal fault

00:45:04.000 --> 00:45:15.000
 normal stress when comparing the Geysers in Oklahoma, we can see that earthquakes at the Geysers are correlated with fault 

00:45:15.000 --> 00:45:19.000
 normal stress in earthquakes in Oklahoma are correlated with fault 

00:45:19.000 --> 00:45:24.000
 shear stress. With these results we can infer that pore pressure response to fault 

00:45:24.000 --> 00:45:27.000
 normal stress is important in Oklahoma  because earthquakes are insensitive to fault normal stress.

00:45:27.000 --> 00:45:33.000
This indicates that Skempton's B-coefficient must be close to one.

00:45:33.000 --> 00:45:37.000
However, at the Geysers earthquakes are very sensitive to fault 

00:45:37.000 --> 00:45:41.000
 normal stress rather than fault shear stress, and the Air Force.

00:45:41.000 --> 00:45:43.000
Kemp's B-coefficient must be low.

00:45:43.000 --> 00:45:58.000
So we have shown that earthquakes at the Geysers in Oklahoma are indeed sensitive to tidal stresses, and we can additionally infer a pearl elastic conditions at the two sites.

00:45:58.000 --> 00:45:59.000
So sorry, just looking at the plot. Thanks, Andy.

00:45:59.000 --> 00:46:13.000
Next up is Tamara Jepson, who's gonna talk about the effective thermal and mechanical processes on hydraulic transmissivity evolution.

00:46:13.000 --> 00:46:14.000
In this presentation I'll be sharing some work we've been doing here at the USGS 

00:46:14.000 --> 00:46:21.000
looking at the evolution of fluid flow in fractures and geothermal systems.

00:46:21.000 --> 00:46:30.000
This work has been funded by the Utah Forge Project, which is an underground laboratory focused on the development testing of technology for enhanced geothermal system.

00:46:30.000 --> 00:46:34.000
While this work that we're talking about is very focused on that, 

00:46:34.000 --> 00:46:42.000
it's also very relevant to any geothermal system where fractions are important into understanding earthquake processes at depth.

00:46:42.000 --> 00:46:44.000
The evolution of fluid flow in geothermal systems 

00:46:44.000 --> 00:46:50.000
 As a result of interacting thermal, mechanical, and chemical processes is not well understood.

00:46:50.000 --> 00:46:58.000
There's quite a bit of laboratory data showing that mechanical processes like shear dilation and sheer enhance compassion can have a significant influence on probability.

00:46:58.000 --> 00:47:03.000
But at higher temperatures most of the work has been done on stationary fractures,

00:47:03.000 --> 00:47:09.000
looking at how probability changes over time as a result of thermal and chemical processes.

00:47:09.000 --> 00:47:13.000
But these tests tend to ignore the effect of shearing,

00:47:13.000 --> 00:47:21.000
 but we know that the interaction of all these processes and their joint effects on the evolution of fluid transfer properties are likely to be very important in

00:47:21.000 --> 00:47:30.000
 understanding earthquake behavior in very relevant and enhanced geothermal systems, where shear stimulation is a common method for increasing productivity.

00:47:30.000 --> 00:47:36.000
We have been conducting laboratory experiments focused on understanding how fluid flow evolves in geothermal systems.

00:47:36.000 --> 00:47:39.000
So we're using a [indiscernible] pressure vessel and our sample material is a Westerly granite. We have one inch 

00:47:39.000 --> 00:47:50.000
 diameter cores that is cut with a sawcut that is roughened to simulate the fracture surface that

00:47:50.000 --> 00:47:57.000
 sample is placed into the pressure vessel, and we apply a constant confining pressure of 30 megapascals (MPa).

00:47:57.000 --> 00:48:00.000
The sample surrounded by a furnace and so we can examine different temperature conditions 

00:48:00.000 --> 00:48:08.000
 and we've gone up to 200°C, and then, of course, we want to saturate it and

00:48:08.000 --> 00:48:12.000
 look at fluid flow. So we have these offset boreholes drilled into the samples 

00:48:12.000 --> 00:48:17.000
 and that allows access to the pore fluid into that fracture and we can apply a pressure.

00:48:17.000 --> 00:48:27.000
We use an average pressure of 10 MPa, leaving us with an effective confining pressure of 20 MPa. To do the flutter tests 

00:48:27.000 --> 00:48:31.000
 we apply a higher pore fluid pressure at the top of the sample than we do at the bottom.

00:48:31.000 --> 00:48:36.000
This difference of pressure causes fluid to flow along the sawcut from one end to the other.

00:48:36.000 --> 00:48:42.000
We can then measure that flow rate and use it to examine how things are changing during the experiment.

00:48:42.000 --> 00:48:45.000
We also want to look at how shear splicing effects fluid flow.

00:48:45.000 --> 00:48:50.000
So we are conducting the flow through tests during slidehold slide experiments.

00:48:50.000 --> 00:49:01.000
So this figure is a schematic diagram of what it a slidehold slide experiment is and where we have the evolution of shear stress with time in this figure.

00:49:01.000 --> 00:49:13.000
So during the first phase of a slidehold slide test, we load the sample along the axis until we start to get slip on that fracture surface.

00:49:13.000 --> 00:49:20.000
During, that time, as we're sliding, we are developing a thin layer of ultrafine gouge 

00:49:20.000 --> 00:49:30.000
due to the fracturing of asparities on the surface, and we slide into a reach of steady state strength, and then we can stop loading.

00:49:30.000 --> 00:49:35.000
Going into the second phase of the slidehold site test, which is the hold period.

00:49:35.000 --> 00:49:40.000
During this time stress is allowed to relax

00:49:40.000 --> 00:49:47.000
 and we can see how fluid flow evolves in kind of this quasi-stationary period.

00:49:47.000 --> 00:49:52.000
We can then start to load the sample again and get continued sharing 

00:49:52.000 --> 00:49:56.000
 going into another slide period.

00:49:56.000 --> 00:50:08.000
Now with the flow rates measure during the flow through tests I calculate the hydraulic transmissivity I'll be dealing with the fracture transmissivity, which is different than aquifer transmissivity.

00:50:08.000 --> 00:50:22.000
Aquifer transmissivity is the product of hydraulic transmissivity in the thickness of the aquifer with units of meters squared per second. Whereas, fracture transmissivity is the product of permeability, in fracture aperture with units of

00:50:22.000 --> 00:50:40.000
 meters cubed. Now, this transmission I'm showing here assuming a rectangular fracture with constant width, we have etched a groove into the socket surface for perpendicular to the boreholes to facilitate parallel flow like we would

00:50:40.000 --> 00:50:46.000
 get in rectangular fracture, but reality is that the flow paths are not completely parallel.

00:50:46.000 --> 00:50:48.000
So we have done some preliminary numerical modeling to provide a correction to the rectangular 

00:50:48.000 --> 00:50:54.000
 full calculation.

00:50:54.000 --> 00:50:57.000
That corrected data is what we're looking at in the coming slides.

00:50:57.000 --> 00:51:09.000
Well, this crash may not be exact. We are mostly looking at changes in the transmissivity in these tests, and so we think it's pretty good for what we're looking at.

00:51:09.000 --> 00:51:10.000
These figures show the evolution of hydraulic transmissivity with experiment duration on both a log 

00:51:10.000 --> 00:51:27.000
 linear and log scale. The 0 time is the time at which we started to shear our fracture and the gray shaded time intervals are the sliding periods where the white intervals are 

00:51:27.000 --> 00:51:32.000
 the hold periods during our slidehold slide experiments.

00:51:32.000 --> 00:51:36.000
Data acquired at room temperature is indicated by the blue symbols,

00:51:36.000 --> 00:51:46.000
100 degrees Celsius is indicated by the black triangles, and 200 degrees by the yellow and orange triangles.

00:51:46.000 --> 00:51:50.000
At all temperatures, there's an overall reduction in hydraulic transmissivity over the course of the experiment.

00:51:50.000 --> 00:52:01.000
Initially, this occurs very rapidly, as the shearing of the fracture surface leads to development of where products and combination and compaction over the resulting gouge layer.

00:52:01.000 --> 00:52:02.000
Once the gouge has developed the rate of hydraulic transmissivity 

00:52:02.000 --> 00:52:08.000
 loss slows.

00:52:08.000 --> 00:52:26.000
In the long-term, loss rate can be quantified by fitting a parallel relationship to the data. When we do this we can see that the rate of decay increases with increasing temperature as is expected for our genius type process.

00:52:26.000 --> 00:52:35.000
Now shearing of the fracture results in these transient increases in hydraulic transmissivity that are superimposed, on the long-term loss rate.

00:52:35.000 --> 00:52:47.000
These increases are caused by shear-dilation when shearing stops hydraulic transmissivity decays rapidly and quickly returns to follow the long-term loss rate.

00:52:47.000 --> 00:52:52.000
In the next slide, we'll focus on the decade rate of these transients.

00:52:52.000 --> 00:53:02.000
Starting with the decay at the start of this 500,000 s hold period highlighted by the red box. To focus on the transient decays in the figure on the left,

00:53:02.000 --> 00:53:08.000
I've subtracted the long-term decay curve from the hydraulic transmissivity data.

00:53:08.000 --> 00:53:18.000
So we're looking at the residual transmissivity at the start of the 500,000 s hold period that was highlighted in the previous slide.

00:53:18.000 --> 00:53:30.000
Now, the reduction hydraulic transmissivity follows an exponential relation of the form shown here, and we see that the decay rate actually decreases with increasing temperature.

00:53:30.000 --> 00:53:41.000
This is opposite of what was observed in the long-term loss rate and is very unexpected as we're not aware of a mechanism that occurs more slowly at higher temperatures at the moment.

00:53:41.000 --> 00:53:46.000
As I went through, and did this for every hold period, which is shown in the figure on the right.

00:53:46.000 --> 00:53:52.000
So here we have the exponential decay with hold duration 

00:53:52.000 --> 00:53:58.000
 with blue symbols indicating room temperature data; black is 100 degrees, and orange is 200 degrees 

00:53:58.000 --> 00:54:11.000
Celsius. We can see that we do very consistently have this negative correlation between the exponential decay rate and temperature.

00:54:11.000 --> 00:54:19.000
I was concerned that the initial formation of our gouge layer could be influencing this, and so I've highlighted data that was acquired during the initial 8 days of 

00:54:19.000 --> 00:54:26.000
 the experiment as these X's, and then the circles indicate data acquired later on in the experiment 

00:54:26.000 --> 00:54:35.000
 once that gougelator has developed fully, and I'll know that the experiments ran for about 3 weeks usually.

00:54:35.000 --> 00:54:49.000
So we have quite a bit of data acquired after the initial 8 days. When we look just at the circles you can really see that negative correlation between decay rate and temperature.

00:54:49.000 --> 00:55:04.000
So it looks like, maybe when the gouge is driven far from equilibrium by sure dilation, we end up with a process as affecting the short-term hydraulic transmission decay that is different from the long-term mechanism.

00:55:04.000 --> 00:55:13.000
This might be something like looking at primary creep versus secondary creep.

00:55:13.000 --> 00:55:20.000
We are very interested in understanding the underlying mechanisms that are responsible for the reserved evolution of hydraulic transmissivity.

00:55:20.000 --> 00:55:28.000
With the short-term decay observed at the start of the holds, we found that there is this negative correlation between the decay rate and temperature.

00:55:28.000 --> 00:55:31.000
Well, don't know what mechanism will result in this relationship 

00:55:31.000 --> 00:55:37.000
 we might be able to learn something by comparing to the initial rapid decay at the start of the experiments.

00:55:37.000 --> 00:55:43.000
Now we know that this is initial decay is tied to development 

00:55:43.000 --> 00:55:50.000
 Comminution and compaction of the gouge layer that we know develops on the surface of the fracture.

00:55:50.000 --> 00:55:56.000
Now the timing of the transition from this rapid decay to the slower long-term decay 

00:55:56.000 --> 00:56:07.000
 that increases with increasing temperature, that transition occurs around 3 hours at room temperature; 7 hours at 100 degrees; and around 15 hours in the 200 degree experiments.

00:56:07.000 --> 00:56:18.000
This suggests that the development of that surface, the development of that gouge layer takes longer at higher temperatures

00:56:18.000 --> 00:56:26.000
 and so it's possible that the rapid decay and the negative dependence on temperature is due to similar mechanisms,

00:56:26.000 --> 00:56:31.000
 possibly mechanical compaction of the gouge until it reaches a mechanical equilibrium 

00:56:31.000 --> 00:56:44.000
 state, and then transitions into the longer-term decay rate. Now that longer-term decay rate may be due to more chemical processes like dissolution and precipitation.

00:56:44.000 --> 00:56:59.000
When we look at the microstructures on the fracture surfaces we see evidence of dissolution in the form of curved grain boundaries in widen fractures as well as possible etching and pitting of some of the mineral surfaces and we also see evidence of

00:56:59.000 --> 00:57:06.000
 precipitation, or we see these fibrous minerals developing on the surfaces of many other grains.

00:57:06.000 --> 00:57:23.000
Possibly this is a clay mineral it appears to have a higher aluminum content which may suggest it's kaolinite and so this may explain explain the longer, slower decay of hydraulic transmissivity in these are processes 

00:57:23.000 --> 00:57:30.000
 we expect to have a positive correlation with temperature, which is what we see in the long-term decay rate.

00:57:30.000 --> 00:57:40.000
The evolution of hydraulic transmissivity is a result of multiple interacting processes when those processes appear to be different when the gouges have been far from equilibrium by active sharing.

00:57:40.000 --> 00:57:41.000
Now, the initial wrapper reductions in hydraulic transmissivity are likely 

00:57:41.000 --> 00:57:54.000
 the result of rapid compaction of the fracture. This appears to occur too quickly to be related to a chemical process, so it's likely due to mechanical processes.

00:57:54.000 --> 00:58:09.000
It looks similar to the primary creep observed in creep experiments on Coulomb material, and as in creep test once this initial phase is over, we switch to a different regime for transmissivity of loss occurs more slowly then the decay rate is positively correlated with

00:58:09.000 --> 00:58:11.000
temperatures.

00:58:11.000 --> 00:58:16.000
So, what are the implications for enhanced geothermal systems and induced seismicity?

00:58:16.000 --> 00:58:23.000
Well, let the majority of transmissivity loss occurs during the initial or primary phase, right after we stop sharing. This primary decay occurs at a slow rate at higher temperatures.

00:58:23.000 --> 00:58:37.000
So that's good news for enhanced geothermal systems and the increases in transmissivity due to shear stimulation should resist longer relative to cooler systems.

00:58:37.000 --> 00:58:45.000
This also suggests that the development of overpressure may be delayed reducing the potential for induced seismicity in geothermal areas.

00:58:45.000 --> 00:58:49.000
But once the primary decay ends. The secondary phase of transmissivity 

00:58:49.000 --> 00:58:51.000
loss occurs more rapidly in geothermal areas in the cooler areas.

00:58:51.000 --> 00:58:57.000
So longer term this may cancel out.

00:58:57.000 --> 00:59:02.000
But what if cold water is being injected at rates and volumes that cause the system to cool?

00:59:02.000 --> 00:59:06.000
We may end up with the worst of both compassion regimes.

00:59:06.000 --> 00:59:20.000
We could see the rapid primary phase of transmissivity loss when the system was cooled, and then, as the system warms up, we could also end up seeing the faster high-temperature secondary phase so it'll be very important to think about the rates 

00:59:20.000 --> 00:59:32.000
 of fluid injection, and how that will affect the system as a whole.

00:59:32.000 --> 00:59:47.000
I think that's it. Thank you very much, Tamara. Last, but certainly not least, is Roland Gritto, who's gonna talk about high-resolution seismic imaging of the Geysers Geothermal Field.

00:59:47.000 --> 00:59:50.000
Good morning everybody, and thank you very much for inviting me to this workshop.

00:59:50.000 --> 01:00:01.000
As you have heard, my talk is on high-resolution seismic imaging at the Geysers Geothermal Reservoir, and this is work I've been doing with my co-office 

01:00:01.000 --> 01:00:23.000
Steve Jarpe (JDS), Craig Ulrich (LBNL), and Craig Hartline (Calpine). I'd like to take the opportunity to thank Calpine Corp. for sharing their three reservoir models specifically Craig Hartline developed his over the past years, and I would like to acknowledge the financial support by the California Energy Commission.

01:00:23.000 --> 01:00:37.000
The objectives of this project were to leverage the development of low-cost seismic recorders, it has been developed over the past years to generated events that worked off a network of 100 stations.

01:00:37.000 --> 01:01:07.000
Then to demonstrate that this dense network, together with automatic processing, allows for high-resolution tomographic imaging of the reservoir properties, and then to demonstrate that these images correlate to rock and fluid-properties. At the bottom right you see a rendering of the geothermal reservoir from Craig Hartline and this was his interpretation of our results.

01:01:09.000 --> 01:01:20.000
So on the left side you see a map of the Geysers, the red polygon denotes the outline of the steam field; and the green triangles show the locations of the permanent seismic network installed at the Geysers.

01:01:20.000 --> 01:01:32.000
The red triangles are the temporary network that we're using in this study.

01:01:32.000 --> 01:01:40.000
It was set up in 2018. Supposed to be for one year, but it's still running although only running a few stations a year.

01:01:40.000 --> 01:01:41.000
Next square is 5 by 5 kilometers study area 

01:01:41.000 --> 01:01:53.000
 that Calpine selected the study because they want to do additional development in this area.

01:01:53.000 --> 01:02:18.000
On the lower right, you see two folders of our stations. On the left you see the inside of the station [indiscernible] battery and the three geophone components below the circuit board. On the right you see how the stations are attached to concrete.

01:02:18.000 --> 01:02:23.000
So, the seismic imaging and seismic data processing is done with different 

01:02:23.000 --> 01:02:33.000
software packages for the detection of P- and S-wave phase arrivals. We use a placement deep-neural-network algorithm and GaMMa,

01:02:33.000 --> 01:02:43.000
and event association algorithm to determine primary locations.

01:02:43.000 --> 01:02:53.000
[cough] Excuse me. The seismic imaging is done using

01:02:53.000 --> 01:02:56.000
 the [indiscernible] S-waves for the locations of the hypocenters.

01:02:56.000 --> 01:03:15.000
It uses the fast eikonal FD solver for fault modeling and therefore is highly efficient for the situation at the Geysers where you have lots of seismicity, and in this case has been used with more than 30,000 events at 150 node spacing.

01:03:15.000 --> 01:03:31.000
In general, the P- and S-wave velocity features you will see are indicative of the large-scale geology, while the Vp/Vs-ratio is more indicative of the small-scale injection and production activities. In the following maps, I'll show you cross-section on maps of these

01:03:31.000 --> 01:03:39.000
properties. So first, I'll show you horizontal maps through the reservoir. On the left side I'll show you the P-

01:03:39.000 --> 01:04:01.000
 wave velocity. On the right side the S-wave velocities from our own version. These are result of maps and if you can read the depth slices [indiscernible] on top of these panels you can read them, but in general it goes from the shadow reservoir over here to deeper reservoir

01:04:01.000 --> 01:04:08.000
 levels and further on until you reach the deepest levels that I'm showing you. The same for the S-wave on the right side.

01:04:08.000 --> 01:04:38.000
So on the left side, the shuttle subsurface, you can see that [indiscernible] very low P- and S-wave velocities, and then, as we go deeper, we see that when we transition to higher velocities and until we see the highest velocities at the bottom of the reservoir. This is actually the [indiscernible] side as we will see later and the mega-intrusion into the reservoir known as the source of the [indiscernible] reservoir.

01:04:38.000 --> 01:04:46.000
The same is true for the S-waves on the right side you can see that low velocities are prevailing at the shallow depth 

01:04:46.000 --> 01:04:52.000
 then as we go to deeper depths we see the higher velocities come through, which again [indiscernible].

01:04:52.000 --> 01:05:02.000
Yet when we look at the Vp/Vs-ratio, we don't see the screen structure. We see 

01:05:02.000 --> 01:05:26.000
 much more heterogeneity, smaller anomalies that seem to be located in different parts of our study, and we get into that in the moment, but that shows already the difference between seismic waves and Vp/Vs velocities ratios.

01:05:26.000 --> 01:05:30.000  
In these slides, I'll show you cross-section through the model on the lower right here you won't see our study area again.

01:05:30.000 --> 01:05:46.000
I'll show you three cross sections, a double prime is shown over here. The double prime here [indiscernible] moment.

01:05:46.000 --> 01:05:55.000
This is the top of the felsite map here and you can see the crest of the felsite coming down in this collection.

01:05:55.000 --> 01:06:14.000
So when we superimpose the information from the 3D reservoir model, you can see these different interfaces and you see that these low velocities on the top part of the reservoir are associated with the caprock in this case it's greenstone.

01:06:14.000 --> 01:06:25.000
then we transition into the reservoir 

01:06:25.000 --> 01:06:33.000 
 we see higher velocities with graywacke higher velocities with hornfelsic graywacke and felsite at the bottom.

01:06:33.000 --> 01:06:37.000
So we can see actually the felsite is characterized as high velocity see it on the bottom. On the top right side 

01:06:37.000 --> 01:06:49.000
Cross section we can seen the cross section [indiscernible]. The top file size is dipping

01:06:49.000 --> 01:06:53.000
[cough] in this direction. 

01:06:53.000 --> 01:07:14.000
Similar results are observed for the S-waves. Then we have low velocities for the caprock and the shallow subsurface [indiscernible] higher 

01:07:14.000 --> 01:07:38.000
velocities graywacke and the hornfelsic graywacke as we go down further into the reservoir until we get to the felsite [indiscernible] velocities down here. Yet again, as I mentioned before, the VpVs-ratio on the right side doesn't show spatial correlation with these interfaces it shows more small-scale heterogeneity indicative of these injections and production activities as I will show you in the next slides.

01:07:38.000 --> 01:07:39.000
So now we transition into the correlation of our results to reservoir properties 

01:07:39.000 --> 01:07:54.000
 and I want to show you reservoir properties results that we observed. Before we get into that I don't want to dwell on this 

01:07:54.000 --> 01:08:17.000
 but Vp/Vs-ratio has been used in the past quite a bit in the oil and gas industry, it's used to distinguish oil from gas deposits but also in geothermal operations and volcanic operations it has been used.

01:08:17.000 --> 01:08:27.000
And you should know if you are not familiar that with high Vp/Vs-ratio are associated with liquid fluids in the subsurface and low Vp/VS-ratio are associated with gases fluids in the subsurfaces such as steam at the Geysers.  So at the Geysers you distinguish between water and steam being high Vp/Vs-ratio and low Vp/Vs-ratio respectively.

01:08:27.000 --> 01:08:57.000
And now I'm showing you the first correlation of Vp/Vs to steam. What we see here is a map at the depth of 700 m, a horizontal map. Okay. Again you see the outline of our study area and we see anomalies here.

01:08:58.000 --> 01:08:59.000
And you can see probably Vp/Vs-ratio scale here of low Vp-Vs anomalies situated here. And we are superimposing boreholes

01:08:59.000 --> 01:09:07.000
these are steam production boreholes. And what you see here.

01:09:07.000 --> 01:09:35.000
Discs are used in the reservoir model as an indicative for how much steam is being produced. The bigger the spear of the disc [indiscernible] the more steam being produced, and you can see here that there is a fairly large disc association with this large anomaly. And on this side here a smaller disc smaller Vp-Vs anomalies. In the next slide you will see cross sections through this

01:09:35.000 --> 01:09:44.000
model, and they are shown here. This is a view from the southeast. This is north over here.

01:09:44.000 --> 01:10:05.000
Here we have a view from the southwest again. North is in this direction. And now we can see the trajectory of the boreholes coming down into this very low Vp/Vs anomaly over here. Here as well, and you can see at least the quarter of this disc because it's sliced by these [indiscernible] diagrams here. 

01:10:05.000 --> 01:10:14.000
It's very nice [indiscernible] anomaly in the surrounding area of this borehole.

01:10:14.000 --> 01:10:34.000
The next correlation, I want to show you is how Vp/Vs correlates with water. And here on the left side, we have a map with the location of Prati 9 printed on top of our results. Shown here now again is the Vp/Vs-ratio.

01:10:34.000 --> 01:10:42.000
 each start is one noted with our inversion [indiscernible]. Prati 9 is one large volume 

01:10:42.000 --> 01:10:50.000
water injector at the Geysers. What we see also here is the trace of the Big Sulfur Creek fault.

01:10:50.000 --> 01:10:57.000
Big Sulfur Creek fault is a sealing fault and 

01:10:57.000 --> 01:11:03.000
 it's impenetrating or it doesn't allow the water to penetrate through the 

01:11:03.000 --> 01:11:22.000
 fault dipping towards the southwest and acts like a boundary for the reservoir towards the southwest, here.

01:11:22.000 --> 01:11:36.000
And what we see is from our Vp/Vs analysis we have the highest Vp/Vs wells in our area just below Prati 9; indicating that the border actually mentioned is right below the injection of Prati 9. And then we see this trail of high Vp/Vs ratio following down here and cooling 

01:11:36.000 --> 01:11:44.000 
against this fault. We interpret this as an area where we have a flow of water through the reservoir 

01:11:44.000 --> 01:11:52.000
 and then it goes [indiscernible] fault. And you can see this on the right side where we have the shear modulus [indiscernible] at this depth.

01:11:52.000 --> 01:12:00.000
You can see the shear modulus down here. [indiscernible]

01:12:00.000 --> 01:12:08.000
Very low shear modulus in the same area here. So low shear modulus is indicative of lightning.

01:12:08.000 --> 01:12:12.000
Fractured block and interpretation is like this is an area that is highly fractured, or more fraction than the rest of the reservoir. [indiscernible]

01:12:12.000 --> 01:12:33.000
Ten years ago, I did a study with data from 2011, and I know it is also very high

01:12:33.000 --> 01:12:42.000
Vp/Vs anomaly in this area from the big water injector that was situated here. Likely also water that was cool against the fault and therefore showed up in our Vp/Vs energy.

01:12:42.000 --> 01:12:48.000
Here I show the shear modulus at different depths at 760 m, and what is shown here in the lower right,

01:12:48.000 --> 01:12:54.000
 Is an area was very high shear modulus.  This is the highest values that we've seen in

01:12:54.000 --> 01:13:03.000
our area. And this correlates to what Calpine refers to as the Dead Zone over here.

01:13:03.000 --> 01:13:10.000
So this is a cross section that's roughly trending along this dash line over here.

01:13:10.000 --> 01:13:15.000
And what you can see is the correlation here with this zone,

01:13:15.000 --> 01:13:21.000
it's this zone in here that's called the Dead Zone. The reason why it's called dead is because it doesn't consist of any

01:13:21.000 --> 01:13:36.000
 steam.   As you can see, these are all borehole trajectories, and the red disk indicate the steam coming into the bore and then you can see all these dry no stream on this area.

01:13:36.000 --> 01:13:59.000
So this was interpreted by Calpine as a very competent gray wiping developed [indiscernible] confirmed by our data, because we see very high shear modulus is indicative of competent work as well. So the last picture I want to show you here is a figure

01:13:59.000 --> 01:14:12.000
 of S-wave velocity superimposed on these faults that have been characterized by Craig Hartline 

01:14:12.000 --> 01:14:25.000
 from seismicity in this area. We can see that jumps in velocity from one area to the other, and they call it very nice with these faults [indiscernible]

01:14:25.000 --> 01:14:32.000
You see, for instance, this area nearby is found by these force and jumps to other sides of the forward show jumps in velocity as well.

01:14:32.000 --> 01:14:45.000
And so this is indicative that with these high resolution dense networks, where it was in high resolution and obtained seismic images where velocity is not necessarily smooth. [indiscernible]

01:14:45.000 --> 01:14:54.000
So in conclusion, we can take home the following messages.

01:14:54.000 --> 01:15:01.000
The availability of cost-effective sensors make operation of dense networks affordable and we have seen this over the past years [indiscernible].

01:15:01.000 --> 01:15:27.000
P- and S-wave velocities are useful to image large-scale reservoir structure. Vp/Vs-ratios for interpretation of injection and production operations and to support the drilling program of the geothermal operators, and in general dense seismic network can provide sufficient

01:15:27.000 --> 01:15:32.000
resolution for detailed imaging of reservoir structure 

01:15:32.000 --> 01:15:36.000
 and faulting.

01:15:36.000 --> 01:15:45.000
And with that I thank you very much and I will take any questions you may have in the formal discussion session.

01:15:45.000 --> 01:15:52.000
A big hand to all the speakers. Excellent job, and really nice presentation on what is going on around the geysers before we start the Q&A

01:15:52.000 --> 01:15:55.000
session in earnest Denise is gonna lead there's a request to play Jared's video.

01:15:55.000 --> 01:16:03.000
that didn't play.

01:16:03.000 --> 01:16:04.000
Jared, can you share your screen and play that video for us?

01:16:04.000 --> 01:16:26.000
Please.

01:16:26.000 --> 01:16:30.000
Yeah. Okay. Sorry. Yeah.

01:16:30.000 --> 01:16:32.000
Perfect. Thank you.

01:16:32.000 --> 01:16:38.000
Oh, no, it's not gonna play

01:16:38.000 --> 01:16:44.000
Hold on!

01:16:44.000 --> 01:16:46.000
Let's see, maybe if there's some questions.

01:16:46.000 --> 01:16:58.000
Now I'll try to figure out how to play a video while streaming.

01:16:58.000 --> 01:17:01.000
Yeah, we can start with a few questions that were already in the chat. [laugh]

01:17:01.000 --> 01:17:12.000
Yeah, we have tech ground ones right now. So this question his, for let me go back to my notes.

01:17:12.000 --> 01:17:27.000
This question for, please? Yes, this is from David, and both the Magar and Shapiro models do they refer to maximum expected magnitude or maximum possible magnitude?

01:17:27.000 --> 01:17:28.000
Yeah. Sorry that I did not reply in the chat directly.

01:17:28.000 --> 01:17:34.000
Do you hear me? Well, by the way.

01:17:34.000 --> 01:17:40.000
Yeah, okay. And I was doing a little bit of research on this.

01:17:40.000 --> 01:17:53.000
And I think aren't in his paper, basically talks about the let me look at it again.

01:17:53.000 --> 01:17:58.000
The maximum earthquake that could be induced by a given fluid injection project.

01:17:58.000 --> 01:18:04.000
So that sounds to me a little bit like possible rather than expected.

01:18:04.000 --> 01:18:14.000
I would say, and I think some some similar statement is done in the in the chapel paper, 2,011 and 2,013.

01:18:14.000 --> 01:18:19.000
So I think it's more in the lines of possible magnitude.

01:18:19.000 --> 01:18:29.000
I think, but in in the case of our our the case of devices or the guys in aria where we were analyzing, I think it's it.

01:18:29.000 --> 01:18:40.000
It's relatively clear why it doesn't fit, because the elastic deformation that is predicted by Magar model is linked with fluid water.

01:18:40.000 --> 01:18:46.000
But in our case a lot of the water is evaporating, so it's not really surprising.

01:18:46.000 --> 01:18:52.000
It's not that it does not fit free

01:18:52.000 --> 01:18:54.000
Great. Thank you.

01:18:54.000 --> 01:18:56.000
And it sounds like we have the video

01:18:56.000 --> 01:19:13.000
I sent Jared's video to John so that John can try play

01:19:13.000 --> 01:19:16.000
Cool thanks.

01:19:16.000 --> 01:19:46.000
Yeah.

01:19:50.000 --> 01:19:55.000
Yeah. Pause. Right around there.

01:19:55.000 --> 01:20:04.000
Cool. So I guess so things to point out so the geysers now are about like 6 or 7 kilometers.

01:20:04.000 --> 01:20:26.000
The seismicity has slowly dissipated. As transition is countered, and then to the north east there's this reduction in seismicity between the Bartlett Springs and the barriers of fall right with that bright red blob is to the northeast

01:20:26.000 --> 01:20:34.000
So you can go ahead and play it again

01:20:34.000 --> 01:20:42.000
And then, if you pause it, you'll start to see to the north east of the lake.

01:20:42.000 --> 01:20:53.000
There's a small seismic cluster, and then also to the southwest there will be a small cluster that will develop, and those are identified as long period earthquakes.

01:20:53.000 --> 01:21:00.000
And they're not sure if they're related to so seismicity or magmatic activity.

01:21:00.000 --> 01:21:05.000
But Roland might have some other information on those

01:21:05.000 --> 01:21:08.000
Go ahead and play.

01:21:08.000 --> 01:21:12.000
So you can see those 2 clusters

01:21:12.000 --> 01:21:32.000
And really that's about it, said Missy

01:21:32.000 --> 01:21:36.000
Cool.

01:21:36.000 --> 01:21:48.000
Any additional questions for Jerry. Now that we saw the video, all its glory

01:21:48.000 --> 01:21:49.000
If there are.

01:21:49.000 --> 01:21:53.000
Raise your hand in the chat. Whatever you prefer.

01:21:53.000 --> 01:21:56.000
That's right.

01:21:56.000 --> 01:21:57.000
Yes, not

01:21:57.000 --> 01:22:05.000
So Jared, I'll ask you to comment on some conversations that we're having happening in the chat during your during your talk.

01:22:05.000 --> 01:22:10.000
So from us, cause the East stepping north South, seismicity trends and younging towards north be indicative of a new connection.

01:22:10.000 --> 01:22:21.000
Shortcut structure forming between the Bartlets spring Salt, and Makomosal.

01:22:21.000 --> 01:22:31.000
This area also coincides with the released, and then on both the Makma faults and the Bartlett Springfield, with the the secondary comment being from seas, you know, agreeing that the release in Ben geometry is an important factor.

01:22:31.000 --> 01:22:40.000
But not fully understood.

01:22:40.000 --> 01:22:45.000
Yeah, that's good question. And I don't completely understand the dynamics of that area.

01:22:45.000 --> 01:23:08.000
But it seems that there is northeastern extension between the Malcolm all and the Bartlett Springs fall, and some have surmised that that's related to movement of the Mendocene or Triple Junction as it moves Northwards, and now there's

01:23:08.000 --> 01:23:23.000
Accommodation there for extension, and whether that was caused by magnetism or or or that is, that's allowing magnetism to occur at the moment

01:23:23.000 --> 01:23:29.000
I guess that's probably what's going on

01:23:29.000 --> 01:23:30.000
Thank you.

01:23:30.000 --> 01:23:33.000
Sure. Did you want to expand on that loss

01:23:33.000 --> 01:23:36.000
Well not to sell the expand. But now that I've seen the video, which is really cool.

01:23:36.000 --> 01:23:41.000
By the way, thanks for putting that all together. Amazing data set and analysis.

01:23:41.000 --> 01:23:50.000
One thing that's you know, sort of suggested to me is seeing that video twice is that at depth like below the really active zone.

01:23:50.000 --> 01:23:59.000
It seems like your contrast is much more concentrated and well defined along a single, let's say, playing.

01:23:59.000 --> 01:24:14.000
But as you come up to the serve towards the service about that horizon, it seems like the the formation, and the structures open up seismicity becomes more complicated, and these are initial on north side, trending seize Mr.

01:24:14.000 --> 01:24:17.000
Patterns become more obvious, so it seems like it's a true going structure.

01:24:17.000 --> 01:24:24.000
There, at this form what I saw, and it's really interesting.

01:24:24.000 --> 01:24:25.000
Thank you.

01:24:25.000 --> 01:24:35.000
Yeah. Thanks.

01:24:35.000 --> 01:24:42.000
Yes, Robert.

01:24:42.000 --> 01:25:00.000
Unmute yourself, Bob.

01:25:00.000 --> 01:25:08.000
Should be a button in the lower left corner of your screen.

01:25:08.000 --> 01:25:11.000
Maybe somebody can unmute it for him.

01:25:11.000 --> 01:25:12.000
Oh!

01:25:12.000 --> 01:25:14.000
Oh, there we go! Thanks!

01:25:14.000 --> 01:25:24.000
Yeah. Jared, it was unclear to me, is my city.

01:25:24.000 --> 01:25:44.000
I see in your plant, and maybe it's just my isight whether that's seismicity associated with the Bartlett Springs faults on? Or is that the fault along the northeast side of the Middle Mountain block

01:25:44.000 --> 01:25:46.000
Yeah, I mean, when it's plotted at the surface, it's pretty close to the Bartlett Springs.

01:25:46.000 --> 01:25:52.000
But as you go further deep it it moves to the north northeast of that fall.

01:25:52.000 --> 01:26:05.000
Okay, the the reason I ask that is because because there is a Amanda Thomas Po has published a couple of papers on cluster seismicity.

01:26:05.000 --> 01:26:12.000
That's over to the southwest, along the east side of Middle Mountain.

01:26:12.000 --> 01:26:32.000
That doesn't appear to be the the faulting doesn't associate with the seismicity doesn't seem to manifest itself as a as a young looking structure at the surface, and it's been suggested, I think that it might be related to magnet.

01:26:32.000 --> 01:26:43.000
Depth. But it's a very controversial, be the cluster that you were pointing to there rather than the Bartlett Springs.

01:26:43.000 --> 01:26:46.000
Which has been known about for a long time, so

01:26:46.000 --> 01:26:57.000
Yeah. Sure. Yeah. Thanks.

01:26:57.000 --> 01:27:03.000
Okay, great. Thank you. We have. We've got some chats for Patricia.

01:27:03.000 --> 01:27:07.000
Yeah, in in the chat, I was wondering if you wanted to comment on it.

01:27:07.000 --> 01:27:14.000
So from Vicki does he flow, provide any constraints on a possible magnetic body to the north that would affect the effect.

01:27:14.000 --> 01:27:21.000
The Bartlett Spring, very exercising city and and the follow-on from from Jared was, you know, good questions.

01:27:21.000 --> 01:27:27.000
The main key flow anomaly at least near the surface, mainly constrained from the geysers to about the late but that's where most of the bill holes are the most used.

01:27:27.000 --> 01:27:44.000
Key pro map. We've developed in 1,992 by the late Mark Walters it may be time for an update the trend of that anomaly is in line with the conductive body that interrupts the size, and city. I don't know if you wanted to go ahead and comment on

01:27:44.000 --> 01:27:49.000
that more.

01:27:49.000 --> 01:28:03.000
Are you asking me? Are you asking that question was actually directed towards Jared's talk about that that very conductive zone that was going off to the north northeast of of the Clear Lake area?

01:28:03.000 --> 01:28:15.000
And so, if there were a magmatic that body there, would you be able to see it in in existing heat, flow, data

01:28:15.000 --> 01:28:19.000
Yeah? Good? Question.

01:28:19.000 --> 01:28:22.000
I don't know. I mean at that point it's probably like 15 to 20 kilometers depth.

01:28:22.000 --> 01:28:32.000
So I guess depends on what's above that in the rocks.

01:28:32.000 --> 01:28:39.000
Whether that would be a conductive, thermally conductive zoom.

01:28:39.000 --> 01:28:47.000
There's not tons of drill holes in that area, because there's

01:28:47.000 --> 01:28:48.000
Yeah, but yeah, I don't. I don't know if you know any differently.

01:28:48.000 --> 01:28:53.000
But

01:28:53.000 --> 01:29:05.000
I was just curious, and I was always wondering whether you might be able to see it in along the Bartlett Springs, for example, any springs that are hot that might be tapping into a deeper warm source.

01:29:05.000 --> 01:29:06.000
But thanks, thanks very much.

01:29:06.000 --> 01:29:21.000
Yeah, yeah, I mean, we've done field work out there in some of the locals I've indicated that there's sulfur in their wells, and some of them are or so.

01:29:21.000 --> 01:29:24.000
Yeah. Some evidence.

01:29:24.000 --> 01:29:33.000
Thanks.

01:29:33.000 --> 01:29:34.000
Thank you.

01:29:34.000 --> 01:29:51.000
Maybe I cannot share with the question from May. So in in one of the cross sections you showed these, deep, long period events, and I mean clearly and I notice that you you show this.

01:29:51.000 --> 01:30:05.000
I couldn't catch it this quickly but it was a yellow, unnormally in resistivity, coming from the east and that's moving through this cloud or pocket of long period events.

01:30:05.000 --> 01:30:11.000
And then coming closer to the surface below the geysers.

01:30:11.000 --> 01:30:19.000
Is, that resolved, or is this just the anomaly or not result?

01:30:19.000 --> 01:30:22.000
Yeah.

01:30:22.000 --> 01:30:26.000
Yeah, so the area long period earthquakes. It's kind of hard to see in cross-section, but in 3D.

01:30:26.000 --> 01:30:38.000
View. It's right at the boundary of the conducted bits in the resistant bits, which is pretty common, is what we see in other volcanic areas.

01:30:38.000 --> 01:30:52.000
Is that you'll see the long periods peripheral to the, to those bodies, and that that conductive zone is pretty well resolved, and it seems that it dips off to the northeast.

01:30:52.000 --> 01:30:53.000
Okay.

01:30:53.000 --> 01:31:11.000
So it deepens to the northeast from the geysers, and not totally sure what that is I mean, if it's because it's like if it's melt above this, as the slab moves to the northeast, if if that's the melt on top of it and it's being

01:31:11.000 --> 01:31:20.000
Pulled back by the crest. I don't really fully understand the dynamics to that yet, but it's pretty well resolved I just don't know what it is

01:31:20.000 --> 01:31:21.000
And you haven't looked at the data yet that you're collecting, or have collected around Q.

01:31:21.000 --> 01:31:29.000
And a we have like a 100 stations all day or something.

01:31:29.000 --> 01:31:30.000
Yeah.

01:31:30.000 --> 01:31:33.000
If you look at that yet it's a part of what you showed already. I'm not here

01:31:33.000 --> 01:31:35.000
Yeah, that's part of what I showed. Yeah.

01:31:35.000 --> 01:31:38.000
Okay.

01:31:38.000 --> 01:31:40.000
Maybe we can talk often more about it.

01:31:40.000 --> 01:31:45.000
Sure.

01:31:45.000 --> 01:31:49.000
Great. Thank you. Hi, Judy.

01:31:49.000 --> 01:31:58.000
Yeah, I have a question for Tamara, and thanks for a really great talk like how you like framed everything and explain the experiment so clearly.

01:31:58.000 --> 01:31:59.000
And it seems like, you know, the development of the gouge seems to be really important.

01:31:59.000 --> 01:32:09.000
And you know I'm sorry if I'm forgetting this from your previous work.

01:32:09.000 --> 01:32:28.000
But I'm curious if you also looked at. You know the evolution of the coefficient of friction, and I guess, thinking of, you know, do the conditions to to stimulate this year, and that the this sheer dilation, if that also changes during the course of your experiments, and if that would have you know

01:32:28.000 --> 01:32:31.000
Impacts in it to your thermal system

01:32:31.000 --> 01:32:37.000
Yeah, so we are looking at how the coefficient friction evolves didn't include that in this talk just for time.

01:32:37.000 --> 01:32:41.000
But we do see some really interesting variations right now.

01:32:41.000 --> 01:32:47.000
I don't know that we see a lot directly correlating with sort of the evolution of Iraq transmissivity in the friction.

01:32:47.000 --> 01:32:55.000
We're actually seeing some weakening behaviors. I've been really unexpected in this very interesting at the moment.

01:32:55.000 --> 01:32:58.000
Not sure. I think kind of initially, I'm looking at.

01:32:58.000 --> 01:33:03.000
I think we're looking at 2 different processes affecting what affecting the friction, and one affecting I draw it.

01:33:03.000 --> 01:33:07.000
Transmissivity, and so there's some interesting things to look at.

01:33:07.000 --> 01:33:08.000
A little bit more. There.

01:33:08.000 --> 01:33:14.000
Hmm cool thanks, and then, you know, I'm also curious with the gouge, you know.

01:33:14.000 --> 01:33:23.000
You can imagine that if the the granite sliding blocks had different mineralogy or green sizes, that that would also have guys an impact of some degree.

01:33:23.000 --> 01:33:30.000
And so I'm curious if you have any plans, to look at different types of kinetic rocks in your experiments.

01:33:30.000 --> 01:33:32.000
Yeah, so we have some plans. Look at some different things.

01:33:32.000 --> 01:33:42.000
We have some actual core samples from the Utah Forge Project, which will provide an interesting contrast so we're still being granitic, but they have some that are very courts rich and some that are courts poor.

01:33:42.000 --> 01:33:45.000
So looking at kind of that very, as well as in the future.

01:33:45.000 --> 01:33:47.000
I'd like to look at. Actually, non-graded.

01:33:47.000 --> 01:33:55.000
I can go with more, maybe some gabros or results and startling sort of subduction zone, Oceanic behaviors. There

01:33:55.000 --> 01:34:05.000
Great thanks.

01:34:05.000 --> 01:34:10.000
Thank you.

01:34:10.000 --> 01:34:23.000
Any other questions, feel free to either unmute yourself or to the question in the chat, and and some verbalizes

01:34:23.000 --> 01:34:29.000
I have a question for anyy. The lori no, we're venturing outside of Northern California.

01:34:29.000 --> 01:34:49.000
But did I understand your results correctly in that 60% of the size Miss City, in Oklahoma is associated with times of titled or increased stresses to the title response on False

01:34:49.000 --> 01:35:04.000
What that number means is is that it's the the so that I have the title function which is normalized to to one essentially.

01:35:04.000 --> 01:35:22.000
And then I I I take the long term rate, and then I superimpose on top of that this other function that is the forcing rate that will either increase or decrease the rate from that long term rate, so that forcing function is 60 the amplitude of that forcing function peak to

01:35:22.000 --> 01:35:27.000
Peak is 60% of the long term rate. So that's what that means.

01:35:27.000 --> 01:35:40.000
Okay, okay, so I did get that wrong. Thanks for the clarification

01:35:40.000 --> 01:35:45.000
Not a question for Tamara, following up a little bit I'm wondering.

01:35:45.000 --> 01:36:08.000
Camera, how you think the the really fascinating results of these fractured properties scale up to a fraction network like at the Kaisers, where there's interactions between fracture sets or whatever that have different temperature and pressure conditions and maybe different apertures or or whatever I was

01:36:08.000 --> 01:36:10.000
just curious to hear your thoughts on that

01:36:10.000 --> 01:36:14.000
Is a really good question, and something that we need to explore more.

01:36:14.000 --> 01:36:26.000
The plan sort of start taking these laboratory results bring them into micro mechanical models of single fractures, and then using that to upscale into sort of the entire fracture network.

01:36:26.000 --> 01:36:33.000
It is gonna get a lot more complex. Complex. We start to look at multiple fractures that can be under slightly different conditions.

01:36:33.000 --> 01:36:39.000
So yeah, it's a good question that I don't think about the answer to. Yet. It's definitely something we're looking into

01:36:39.000 --> 01:36:44.000
Yeah, it's really challenging. Obviously, thanks. It's cool step

01:36:44.000 --> 01:36:51.000
Just a comment on that. It's not just yeah, different fractures and compartments as well.

01:36:51.000 --> 01:37:00.000
Oh, great hardline can speak more to that. But yes, in this while we talk more about this in the past.

01:37:00.000 --> 01:37:01.000
But sometimes you have to wealth next to each other, there's a fall in between when they don't correspond at all.

01:37:01.000 --> 01:37:26.000
Hydraulically, and then, if you go alongside of these falls, you might have correspond several 100 meters of kilometers further in one direction. So it's it's very difficult to quantify in one area

01:37:26.000 --> 01:37:37.000
It's an error. The whole career

01:37:37.000 --> 01:37:39.000
Stop.

01:37:39.000 --> 01:37:43.000
I have a another question for Roland Grido.

01:37:43.000 --> 01:37:52.000
In your Illinoisations there, I'm a framework G, all just so. I'm I'm interested in in in the fault thing that that you are showing in your illustration.

01:37:52.000 --> 01:38:17.000
So. So you show the the Big Sulfur Creek fault with South with thrust teeth on it, implying that it's a southwest dipping fault, and I think you mentioned that those by the thrust teeth are do I assume then

01:38:17.000 --> 01:38:27.000
That that you this is a interpreted, or that it's known that fall is known to be shallow dipping?

01:38:27.000 --> 01:38:33.000
Or is it a, or you just mean that it's a steep southwest dipping fault?

01:38:33.000 --> 01:38:40.000
The reason I ask this is because of the Mercuryville fault zone, which is at least those of us that worked in the steam field years ago.

01:38:40.000 --> 01:39:01.000
Considered to be kind of the bounding southwest fall of the reservoir, and if this is a major Southwest dipping thrust, it would interface with the Mercuryville fault zone to the southwest.

01:39:01.000 --> 01:39:03.000
It would seem so

01:39:03.000 --> 01:39:07.000
Yeah, I, I mean, I'm a seismologist.

01:39:07.000 --> 01:39:12.000
So forgive me if I just indicate some. Some different signs on there.

01:39:12.000 --> 01:39:16.000
I I had a cross section as well. It's a sleep tipping fault.

01:39:16.000 --> 01:39:17.000
Okay, okay, that that makes more sense to me.

01:39:17.000 --> 01:39:32.000
It's different to the yeah. And if you go in the north Westernly extension, then you get the so it's it's kind of an extension to the southeast of it as a Mercury girl.

01:39:32.000 --> 01:39:33.000
Okay.

01:39:33.000 --> 01:39:39.000
If you want to call it's it's actually called the South a quick fault, because it's running along the the creek, you know.

01:39:39.000 --> 01:39:40.000
Yeah, yeah, I, I think.

01:39:40.000 --> 01:39:54.000
Hold itself. Almost the depression weather quick is running, so to speak. And you can actually see software deposit exposed on the hanging wall there as well

01:39:54.000 --> 01:39:55.000
So.

01:39:55.000 --> 01:39:57.000
Okay, yeah, they, there's yeah. There's a series of of stepping steep dipping faults right along the creek that you look pretty young.

01:39:57.000 --> 01:40:06.000
And I probably called that the big silver Creek fault so, or somebody in the department in Dog.

01:40:06.000 --> 01:40:07.000
But okay, thanks for all

01:40:07.000 --> 01:40:11.000
Right? And then, yeah, Jared, not, Jerry.

01:40:11.000 --> 01:40:17.000
I forgot. Oh, I shouldn't know call it from Berkeley.

01:40:17.000 --> 01:40:21.000
Ask me the question whether we have can do this in Fourd and my requires.

01:40:21.000 --> 01:40:22.000
We have done this in fortune. So we did a data set.

01:40:22.000 --> 01:40:30.000
We analyze data sets from 2,005 to 2,012 for 11 main company.

01:40:30.000 --> 01:40:43.000
The differences, and it turns out that we see well the Pbs ratio changes positive changes right next to the Software Creek forward.

01:40:43.000 --> 01:40:44.000
And I mentioned that briefly in my position there is one of the biggest.

01:40:44.000 --> 01:40:55.000
What injectors was located just on the reservoir side of the phone, and it looked like a big part of it.

01:40:55.000 --> 01:41:00.000
Great, what's created there, because the the phone is non penetrating.

01:41:00.000 --> 01:41:09.000
And so we could see that in the Vpds ratio all this time exchange was going to very high Vpds.

01:41:09.000 --> 01:41:12.000
So there there's more indication for that.

01:41:12.000 --> 01:41:28.000
Great hard line, and it's reservoir, and it's it's considered a pounding fall to the reservoir. It's very low steam and rough place on the other side

01:41:28.000 --> 01:41:35.000
Yeah, okay, thank you.

01:41:35.000 --> 01:41:40.000
And will. And just a follow-up question is that in a publication somewhere, or is that they'll be working

01:41:40.000 --> 01:41:49.000
Yeah, this is currently in revision. There was a special issue in J of physics on application, in geothermal. Hey?

01:41:49.000 --> 01:41:59.000
That's part of it.

01:41:59.000 --> 01:42:00.000
I think

01:42:00.000 --> 01:42:02.000
And then one other question before we break you. In a minute or so.

01:42:02.000 --> 01:42:06.000
There's a question for Patty and Roland Briggman.

01:42:06.000 --> 01:42:07.000
Sort of answered. Maybe Roland can unmute and give us the the 2 sentence answer.

01:42:07.000 --> 01:42:20.000
But any results from the Ins or GPS work that indicate a size and deformation. The guys are

01:42:20.000 --> 01:42:39.000
So Mike gave a nice overview of Tps measurements both in earlier work and also he and Gareth have 3 continuous GPS stations at the cases and then Don Vasco has an insert paper from a few years ago in Grl and I put a link.

01:42:39.000 --> 01:42:53.000
To a paper by Chi Hong Liu, who was a a visitor with a pretty, nice, comprehensive, and so analysis from a number of data sets, so that might be useful.

01:42:53.000 --> 01:42:56.000
So when Roland Quito just mentioned time dependent?

01:42:56.000 --> 01:43:04.000
You know, subsurface information. Obviously, that's exciting to people who look at time dependent surface deformations.

01:43:04.000 --> 01:43:09.000
So that'll be interesting to compare

01:43:09.000 --> 01:43:15.000
Thank you.

01:43:15.000 --> 01:43:26.000
Okay. I think we're at the 1115 mark. So thanks to all the speakers for really really interesting presentations and a great discussion both in the chat during the talks and afterwards we're supposed to break right now.

01:43:26.000 --> 01:43:33.000
But we'll keep this open in case people have more questions and want to chat more hopefully.

01:43:33.000 --> 01:43:37.000
The speakers can, you know, give us a couple more minutes here or there?

01:43:37.000 --> 01:43:42.000
But thanks again. Great session

01:43:42.000 --> 01:43:48.000
And thank you, not just to our wonderful speakers. But you are wonderful moderators as well

01:43:48.000 --> 01:43:52.000
Thank you.