Development by Design in West Texas:
Mitigating Energy Sprawl Through Cooperative Landscape Planning
Kei Sochi, Jon Paul Pierre, Louis Harveson, Patricia Moody Harveson, David V. Iannelli,
John Karges, Billy Tarrant, Melinda Taylor, Michael H. Young and Joseph Kiesecker
May 2021
3. A Conservation Vision in the Context of Future Energy Development
Identifying and mapping important landscape values is a critical first step to understanding and managing the possible impacts of future energy development [19, 21]. These values are intended to capture what makes the Tri-County Region unique and important to those who live there, to the state of Texas and to those who visit. Taken together, these values articulate a conservation vision to inform and assess potential impact from future development. The goal is to maintain these values even as the landscape; the communities and economies that exist within it; and the policies that govern them shift and evolve.
Ultimately, long-term conservation success will be inextricably tied to coordinated action among a diverse community of stakeholders. A shared conservation vision and its potential overlap with modeled future energy projections is vital to proactively consider possible conflicts, clarify tradeoffs and strategize how to work in concert to guide an emerging story of development toward a more low-impact future in this rugged and remote West Texas country. Here, RBB presents an overview of methods (for detailed methods, please see supplement sections S3, S4, S5).
Methods: Develop a Landscape Conservation Vision
RBB developed this project’s conservation vision by translating the ecological and social values identified by the SAG into spatial data proxies (Table 1, Figure 2) and convening a science subcommittee of the SAG to review the representations of the biological values. The team adopted a coarse-filter/fine-filter approach to manage the complex organization of biological systems present and the practical limits of existing data representing those systems. The rationale behind the approach is that a first-step focus on terrestrial ecosystems (the coarse-filter) will also conserve the majority of individual species and intact habitats that can support processes needed for the long-term persistence of those species [22]. As a second step, a subset of species was targeted (the fine-filter) that would not be well-represented when attending to a coarse-filter systems approach alone. Candidates may include species that are rare, those with highly specific habitat requirements and/or changing habitat needs over their life cycle, or species that are migratory over long distances [22]. Three species were selected – pronghorn, bighorn sheep and mountain lion – to serve as indicators of important habitat (grasslands and montane forests of the sky islands) and functional connectivity between habitat patches.
All input values were equally weighted and summed together into a single metric representing a surface of cumulative values across the entire landscape. To derive this layer, each input was re-scaled from 0 to 1 to make disparate values comparable. The focal mean value was calculated for each pixel within a 0.6 mile (1 km) radius for each 30 meter grid cell across the study area. This essentially turned every layer into a measure of density. Focal means were not calculated for values already measured as a continuous metric (e.g., irradiance values for nighttime lights). All inputs were resampled to 30m rasters.
For ease in reporting and summarizing, the continuous nature of the cumulative values map were grouped into five equal-interval classes. These classes were titled: very high, high, moderate, low and very low.
| Values | Spatial Proxy | ||
|---|---|---|---|
| (1) | Ranching heritage & private property rights | (1) | Intact landscapes |
| (2) | Grasslands | (2) | Grasslands |
| Water resources | (3) | Riparian areas & wetlands | |
| (4) | Springs | ||
| (3) | Sky islands, wildlife & migratory corridors | (5) | Bighorn sheep model |
| (6) | Pronghorn model | ||
| (7) | Mountain lion model | ||
| (4) | Tourism & hunting | (8) | Pronghorn herd units |
| (9) | Mule deer herd units | ||
| (10) | Recreational routes & trails | ||
| (11) | Managed areas | ||
| (5) | Viewsheds | (12) | Viewsheds |
| Dark skies | (13) | Dark skies | |
| Remoteness | |||
| (6) | Community, safety & quality of life | ||
| (7) | Culture, music & arts | ||
Table 1. Landscape values identified by the Stakeholder Advisory Group and the associated spatial proxies mapped (+ values were not mapped).
Figure 2. Tri-county values mapped across the 18-county study region (1-13).
Methods: Project Future Potential for Energy Development
The goal in estimating future energy development patterns is to determine the likelihood (probability) that oil and gas, solar and/or wind development may occur at a particular location over the next 30 years, circa 2050. To the extent possible, a similar approach was tailored for each energy system by following these key steps:
Map existing infrastructure and evaluate landscape alteration patterns;
Exclude anthropogenic areas restricted from development (e.g., airports, roads, cities, already-developed areas);
Evaluate resource potential (e.g., reservoir quality, solar radiance, wind speed);
Project amount of energy production (i.e., barrel of oil equivalent (BOE) and megawatts (MW)) under different development scenarios, defined in relation to the expected landscape impact as low, medium or high;
Evaluate probability for infrastructure placement at a particular location, based on several criteria, including distance to existing roads, pipelines, transmission lines, and other operating facilities; and
Estimate locations of new production facilities using projected energy production and placement probabilities.
To cover the range of uncertainty that accompanies any forecast, three impact scenarios were identified for oil and gas, solar and wind energy. “Impact” refers to the construction of well pads, pipeline right-of-ways (ROWs), photovoltaic solar installations, etc., all of which facilitate energy production.
For oil and gas development, it is assumed that the trends in landscape change and energy production seen in the last decade will continue into the future unchanged. This represents the medium impact scenario, what RBB is calling the “business as usual” (BAU) case. To help develop the low and high impact scenarios, the number of wells built on each drilling pad was used as a proxy. In the medium scenario, three wells would be built per pad on average, an assumption consistent with observations today, reports found in the literature and in discussion with operators [12, 23, 24]. For the high impact scenario, only one well would be developed on each drilling pad following historical practices (closer to 1.15 wells per pad) [25]. The high impact scenario requires more new pads to be built on the landscape for the same number of wells. The low impact scenario assumed a combination of reduced well-drilling activity given low oil and gas prices, an increased number of wells per pad as a result of technological advances and/or some other economic incentives. Operationally, the low impact scenario would triple the number of wells per pad (compared to the BAU) to nine wells per pad (hence, 1/3 the number of well pads would be needed). Specific placement of new well pads for oil and gas development used the results reported by Pierre et al. (2020) [12].
For solar and wind energy production, future development was estimated together in a single Capacity Expansion Model known as “Switch” [26] to take advantage of differences in peak performance between day (solar) and evening (wind) time periods. Switch matches electricity demand anticipated by the Electric Reliability Council of Texas (ERCOT) over the next 30 years, with future generation capacity for low, medium (i.e., BAU) and high supply scenarios. The Switch model included facilities across the entire ERCOT service area of Texas, which reduces uncertainty that would be introduced by modeling the Trans-Pecos region in isolation (electricity moves across transmission lines into and out of West Texas, but not across state lines). The outcome of this model provided future estimates of both solar and wind generation measured in gigawatts (GW) for each ERCOT region and ultimately for each county in Texas.
Similar to how Pierre et al. (2020) projected oil and gas well pads, a probability raster using expert-driven knowledge of placement criteria was created; however, considering the vast area of land suitable for both wind and solar energy development, we have not placed new facilities on the landscape. Ongoing work will focus on identifying factors important for mapping where new infrastructure is likely to be placed. These factors will be similar to those used for placing hypothetical oil and gas facilities. However, additional factors may be considered when locating electricity generation facilities on the landscape. For example, developers may want to avoid clustering of facilities [12, 27, 28] to limit the potential for transmission line congestion that occurs when too many facilities are built close to one another, leading to curtailment of electricity production. We used the Switch model results to estimate how much new solar and wind electricity production is needed to support projected future electricity demand and then estimated the footprint of future renewable facilities using the median capacity and the median footprint of existing facilities within the ERCOT region.
Results
Cumulative Values Map
Figure 3. Cumulative Values map based on Stakeholder Advisory Group recommendations. This map represents aggregations across the larger 18-county study area of values identified primarily for the Tri-County Region.
Within the 18-county study region, there are ~11 million acres (44,000 km2) in the High or Very High cumulative classes, 40 percent (or 4.4 million acres/17,800 km2) of which is in the Trans-Pecos Region. This is in part because Brewster and Presidio Counties are among the largest counties in the study area (Figure 3). Brewster, Presidio and Jeff Davis Counties are home to some of the largest and most intact occurrences of these high value areas found in the study region. The majority of intact grassland landscapes and movement corridors for our focal species in the highest value classes are found in the Tri-County Region. All told, 81 percent of the Tri-County Region is in the High and Very High cumulative value classes.
Other counties in the study region with large areas in the High and Very High cumulative classes include Hudspeth (1.6 million acres/6,629 km2), Culberson (1.4 million acres/5,701 km2) and Pecos (752,994 acres/3,047 km2). Counties without areas included in the High and Very High classes include Crane, Ector, Loving, Midland, Ward and Winkler (coincidentally, counties with the highest concentrations of energy development) (Figure 4).
Figure 4. Distribution of individual values by county (A) and distribution (in cares) of cumulative value classes by county (B).
Individual Input Values & Irreplaceability
Most of the individual mapped values are well-represented in the High and Very High cumulative value areas (Figure 6). However, special attention needs to continue to be paid to several individual values with outsized importance. Among them are riparian, wetland and spring areas, which are especially critical and irreplaceable for maintaining healthy ecosystems and the species dependent on them in this arid region. These features are not apparent in the cumulative values map, because they are typically mapped as small areas (e.g., springs) or narrow linear features (e.g., riparian areas) and their distribution is limited across the region.
This approach emphasizes areas of high aggregations of individual values. Areas of low cumulative values may still include the presence of otherwise irreplaceable or other valuable elements in good condition. Attention should continue to be paid to understand the individual values present in any area of interest.
Figure 6. Distribution of individual values in each cumulative value class (in million acres).
Energy projections
Oil and gas: Currently, there are an estimated 122,433 well pads in the study area, constituting a total direct footprint of 187,410 acres (758 km2). Our forecasts suggest that an additional 24,925 to 180,849 new well pads could be built by 2050, resulting in an estimated 258,758 to 900,062 acres (1,156-3,642 km2 ) of cumulative direct impact (Figure 5 A1-A3). The bulk of these impacts will be concentrated in Midland, Pecos, Reeves, and to a lesser extent, Culberson Counties.
Figure 5. Low, Medium and High land alteration scenarios for oil and gas development (A1-A3) and suitable lands for potential siting of solar (B) and wind (C) development.
Solar: To date, ~8,896 acres (36 km2) of landscape alteration has occurred from utility-scale solar development in the 18-county study area. We mapped a total of 12 million acres (49,349 km2) of suitable lands with sufficient resource potential for siting future solar facilities (Figure 5B). The “Switch” model calls for an additional 13 to 16.5 GW of capacity from utility-scale solar facilities to be built in ERCOT Region 2 across the different scenarios over the next 30 years. Considering the median capacity (78MW) and size (642 acres/2.6 km2) of facilities in Region 2, an additional 170 to 213 new facilities could be built, resulting in a cumulative direct footprint of between 116,896 and 144,396 acres (473-584 km2). In all three scenarios, we found that the Tri-County Region could expect to host new solar facilities.
| Oil & Gas | Solar | Wind | |||
|---|---|---|---|---|---|
| Current footprint | acres | 187,410 | 8,896 | 118,611 | |
| (km²) | -758 | -36 | -480 | ||
| Possible additional footprint | Low scenario | acres | 98,348 | 75,614 | 270,580 |
| (km²) | -398 | -306 | -1,095 | ||
| Medium scenario | acres | 291,584 | 78,579 | 1,600,005 | |
| (km²) | -1,180 | -318 | -6,475 | ||
| High scenario | acres | 712,652 | 96,371 | 2,741,136 | |
| (km²) | -2,884 | -390 | -11,093 | ||
| Cumulative footprint (current + additional) | Low scenario | acres | 285,758 | 84,510 | 389,191 |
| (km²) | -1,156 | -342 | -1,575 | ||
| Medium scenario | acres | 478,994 | 87,475 | 1,718,616 | |
| (km²) | -1,938 | -354 | -6,955 | ||
| High scenario | acres | 900,062 | 105,267 | 2,859,747 | |
| (km²) | -3,642 | -426 | -11,573 | ||
| Table 2. Current, projected potential future direct land alteration (in acres, km²),and cumulative footprints (current + future) by energy type for entire 18-county RBB areas. | |||||
Wind: To date, 118,611 acres (480 km2) of land have been altered from wind energy development in the 18-county study area. We mapped a total of 13 million acres (54,238 km2) of suitable lands with sufficient resource potential for siting future wind facilities (Figure 5C). The median capacity of existing facilities is ~145 MW, requiring a median land area of ~6,672 acres (27 km2) or a footprint of ~46 acres/MW. When looking to the future, the “Switch” model calls for a capacity increase of 1,628 MW from utility-scale wind farm facilities. Depending on the energy demand scenario (low, medium, high), this would lead to between 23 and 233 new facilities, resulting in a cumulative land alteration ranging between 389,191 and 2.9 million acres (1,575–11,573 km2). See Table 2 for a summation of expected alteration from all energy sources, broken down by energy type (columns) and by scenario (rows).
Using current trends and accounting for the transition of the Texas electricity grid toward renewable energy, it is anticipated that both oil and gas and renewable infrastructure will continue to be installed in West Texas although the Tri-County Region does not appear to be a top priority for any energy type. The likelihood of intensive oil and gas development in the Tri-County Region is low, although future technologies may change the ability to recover hydrocarbons from geologic units in these counties, just as hydraulic fracturing technology unlocked oil and gas reserves elsewhere in West Texas and elsewhere. Solar irradiance levels in the RBB area of West Texas are very high (increasing westward), and wind energy potential is higher eastward and northward; thus, the area is favorable for solar and/or wind energy development, increasing the potential for broader-scale investment, construction and land alteration.
Impacts and Tradeoffs to the Landscape Conservation Vision from Future Potential Energy Development
Potential conflicts with oil and gas: Oil and gas resources are highest in Pecos, Culberson and Midland Counties, but potential overlap with High and Very High value classes are relatively low in Midland County – mostly because Midland is already highly impacted and has fewer areas in these two classes. Depending on the development scenario, ~2,980-23,000 acres (12-94 km2) of High and Very High value areas in Culberson, ~2,040-14,171 acres (8-57 km2) in Pecos and ~1.798-13.125 acres (7.3-53 km2) in Reeves Counties remain vulnerable to conversion (Table 3, Figure 7).
Figure 7. Overlap of Cumulative Value classes and potential oil and gas footprint by county and by development scenario (low, medium, high) (in acres2).
At every development class, most projected potential impacts are confined to areas overlapping the Low and Very Low conservation value classes, but 25 percent of the forecasted development footprint in every scenario reaches into areas in the upper half of the conservation values range. It is anticipated that development is unlikely to be restricted exclusively to the lowest conservation value classes, because the siting of oil and gas well pads is more tightly tied to the location of the hydrocarbon-bearing resource: the oil and gas or shale formation from which the resource can be extracted.
Potential conflicts with renewables: In contrast to the projections for oil and gas development, siting potential for renewable development was evaluated across the entire 18-county study area and the overlap of all potentially suitable sites with conservation values was also evaluated. This overlap significantly overestimates the potential for conflict because of the very low probability that all the areas suitable for renewables will be developed. For example, we estimate ~270,580 to 2,741,136 acres (1,095-11,093 km2) of potential impact from wind energy development across the study area, all of which could be sited within the 8-83% of the 3.3 million acres (13,398 km2) of land area classified with Very Low cumulative value that also has adequate wind resources. Similarly, for solar development, it is projected that the ~84,510-105,500 acres (342-426 km2) of new potential impacts could be sited within 2 to 3 percent of the 3.9 million acres (16,061 km2) of Very Low cumulative conservation value areas that also are suitable for solar facility siting.
The High and Very High cumulative conservation value classes potentially at risk from renewable energy development are concentrated in Culberson, Brewster, Hudspeth, Jeff Davis and Pecos Counties and to a lesser extent, Reeves and Terrell counties (Table 4, Figure 8). The Very High cumulative value areas in Presidio County are especially at risk (solar: 131,200 acres, wind: 87,680 acres). Culberson and Pecos counties are also of particular concern because they could also be under high pressure from forecasted oil and gas development.
Figure 8. Overlap of Cumulative Value classes and renewable resource (solar and wind) potential by county (in acres and km2).
SUMMARY
The Tri-County Region is home to most of the pristine, intact landscape values in the 18-county study area.
Using data about the region’s geology, topography, weather patterns and projections about Texas’ future energy demands and the ongoing transition toward more renewable sources of energy, RBB anticipates that both oil and gas and renewable infrastructure development will continue in West Texas.
Oil and Gas: It is expected that development will be highest in Pecos, Culberson and Midland Counties. Potential conflicts with areas of Very High and High cumulative value classes are most salient in Culberson, Pecos and Reeves Counties.
Renewables: Resource potential for solar facility siting is highest in Pecos, Reeves, Hudspeth, Culberson and Winkler Counties. Potential conflicts with areas of Very High and High cumulative values classes are notable in Hudspeth, Culberson, Presidio and Brewster Counties. Resource potential for wind energy facilities is highest in Pecos, Culberson, Crockett, Val Verde, Terrell and Winkler counties. Potential conflicts with areas of Very High and High cumulative values classes are focused in Culberson, Pecos, Hudspeth and Presidio counties.
The ability to manage siting impacts will vary by energy type: Though this analysis does not include all of the potential impacts associated with oil and gas development (for example, water and air pollution are not discussed), the spatial impacts of forecasted solar and wind development are projected to exceed the land footprint of future oil and gas development [29, 30].
Moreover, the distinct ways that oil and gas, solar and wind energy resources are distributed within the 18-county study region have implications for siting decisions. Specifically, oil and gas production activity is more constrained within the study area, whereas solar and wind resources for electricity generation are broadly located. The larger land area with suitable solar and wind resources across the 18-county study region when compared with oil and gas resources, lends some greater flexibility to siting renewable energy facilities away from the most important areas on the landscape (i.e., areas of High or Very High aggregations of conservation values) to areas with fewer conflicts. This potential siting flexibility at broader scales is less obviously available to oil and gas operators. Prospects for easy tradeoffs diminish further in those counties where future oil and gas development is expected and where areas of High or Very High aggregated values are located.