Mid-South Synergy Electric Cooperative's mostly rural 1,635-sq mile (4,235-sq km) service territory spans six Texas counties: Brazos, Grimes, Madison, Montgomery, Madison, Waller and Walker. Being a rural electric cooperative brings many challenges with respect to vegetation management and vegetation-related outages. Most of Mid-South Synergy's outages in any given year are lightning and vegetation related. While lightning is hard to predict and impossible to prevent, the growth of vegetation can be controlled and its contact with distribution assets can be minimized. The majority of Mid-South Synergy's vegetation-related outages are because of trees growing outside the utility's 20-ft (6-m) right-of-way (ROW).
Studies have consistently shown only up to 15% of tree-related outages are caused by ROW growth. Hazard trees or trees located outside the ROW are the majority of the problem. When one considers the Sam Houston National Forest as an example, where the predominantly pine trees are at least 100 ft (31 m) tall, no amount of conventional clearing within the ROW can prevent damage resulting from trees falling on power lines outside the ROW. It just takes the right weather conditions (high winds or heavy rain) or tree mortality to create an event.
The Lay of the Land
In a year like 2011, in which Texas experienced not only a severe drought but rampant forest fires, tree mortality was at a seldom-seen peak. Dead trees near lines must be removed immediately. This is mainly because decay organisms attack them, weakening their stems, resulting in large limbs and the top breaking off the crown followed by the collapse of the whole tree. Mid-South Synergy is aware of the risk attributed to dead trees in its territory, and this prompted the coop to revisit its vegetation management plan to include a comprehensive plan for cutting down dead trees.
The single most important factor affecting tree growth is soil moisture. According to the State Soil Geographic (STATSGO) database, Mid-South Synergy has 21 soil types in its territory, and each soil type has unique properties. For example, soils differ in their ability to store moisture or retain water. This is referred to as the soil's available water capacity (AWC) and is linked to variable tree growth and mortality. Soil data, specifically the AWC, is a good predictor of high-risk areas where the effects of long periods of drought will be first experienced. In addition to soil variability, there also is a 7-inch (178-mm) west-to-east gradient in annual rainfall across the service territory. Vegetation cover types also are variable, with pine hardwoods dominating most of the territory.
Current Hazard Tree Removal
The typical hazard tree removal sequence starts with a customer call. This prompts the creation of a service order for tree crews. Alternately, the ROW crews identify and cut down hazard trees as they work their regular shifts. Finally, when service crews responding to an outage call identify a hazard tree, they cut or trim the tree depending on the situation. All dead tree data received from these sources are submitted to the geographic information system (GIS) department once the work is completed. The vegetation management geodatabase is then updated using Clearion software. Each dead tree record in the GIS is attributed with the tree species and tree condition (dead or green).
Drought conditions enhance tree mortality. To better prepare for this, GIS was used to analyze hazard tree removal data for 2011, STATSGO soil data and vegetation cover type data. A dead tree area identification model was created in GIS to help in resource allocation for spotting and cutting down dead trees well before a customer calls or, even better, before an outage occurs.
Pines constituted the highest number (75%) of cut dead trees in 2011, followed by oaks, while the remainder consisted of sweet gum, elms and others. The pine dominance in the coop's data could simply be explained by their being the dominant vegetation cover type in the utility's territory. However, pines are also known to be less drought tolerant — except for the Japanese black pine species — compared to most of the oak species.
Based on the STATSGO data, most of the dead trees were cut in the soil mapping unit ID (MUID) TX140 followed by TX179. These mapping units — Depcor-Fetzer-Boy and Frelsburg-Latium-Crockett, respectively — are characterized by high permeability, high drainage, low organic matter content and low clay content, characteristics that, combined, lead to low water retention or low AWC. It also was determined the majority of dead tree-related outages in 2011 were in the soil MUID TX140 and under the pine vegetation cover type.
A field visit to randomly sample the location of dead trees showed that soil type played a much larger role in tree mortality as expected (soil moisture is the single most factor affecting tree condition). In the GIS model formulation, a weight of 0.9 was applied to soil type and 0.1 was applied to vegetation cover type. This resulted in a model skewed toward soil type.
Weighted Overlay Analysis
The following steps were followed in coming up with the GIS model for dead tree risk areas:
A numerical evaluation scale of 1 to 4 was chosen, where 1 is the highest risk and 4 is the lowest risk.
The cell values for the soil type layer and vegetation cover type layer were assigned values from the evaluation. Soil layer TX140 was assigned a value of 1; TX179 a value of 2; TX188, TX205 and TX109 a value of 3; and all others a value of 4. The weight for each soil type was based mainly on the number of dead trees cut in that unit. For the vegetation cover type layer, pine hardwood was given a value of 1; other — a mosaic of many types — a value of 2; post oak woods, forest and grassland mosaic a value of 3; and all others a value of 4.
The soil type was given a 90% influence weight and the vegetation cover type was given a 10% influence weight. Each of the two inputs was multiplied by the weight.
The resulting cell values were added together to produce the risk allocation for dead trees in the Mid-South Synergy service territory.
Most of the dead tree-related outages were prevalent in the soil mapping unit TX140 followed by TX179, TX188 and TX109. TX626 also was seen with a few more outages compared to the remainder of the soil units.
Most of the dead tree outages were reported in the pine hardwood vegetation type. The vegetation type of “other” could not really be classified as just one dominant vegetation type. The purple is the post oak woods, forest and grasslands mosaic.
GIS Model Results
All grid cells with a value of 1 were labeled as the highest risk areas or the most susceptible to dead trees, while areas with a grid cell value of 4 are the least susceptible. The weighted overlay output grid was reclassified from its original five classes to just two classes depicting high- and medium-risk areas.
The next step involved converting the resultant output grid to a shape file, so a spatial join with primary conductors could be carried out. The spatial join allowed for the assignment of susceptibility values to primary conductors based on the polygon in which they fell. Based on the spatial join, it was possible to determine, at the feeder level, the length under each of the two classes. The classes were ranked by percent for each feeder, so if the majority of a feeder had a grid value of 1, then the feeder would be classified as very high risk and so forth. This enabled a grouping of feeders based on dead tree risk. This was done in Microsoft Excel using pivot tables. A table was created listing feeders and their dead tree risk, and this was given to the operations department for deployment of crews to the areas needing the most immediate attention.
The 50 feeders (circuits) were each given a value for dead tree intensity. With this new information, the work flow for taking care of hazard trees was modified, enabling the coop to not just rely on customer calls (reactionary) but be more proactive.
In the new work flow, the GIS department allocates work packets for taking down dead trees based on the feeder susceptibility. This has resulted in intensifying dead tree work more than threefold, thus avoiding many potential outages. For example, close to 3,000 trees were cut during all of 2011, whereas in the first half of 2012, about 15,000 dead trees were removed from the system. Without GIS, the coop would still be relying on customer calls and random scouting to know where hazard trees are located.
A Great Result
This is an example of low-hanging fruit that GIS can help to expose. A few spatial analyses were run to come up with the hazard tree management plan that has since been implemented. With this new process, work assignments are much more efficient and the hazard tree program has already managed to remove trees from the system that would normally result in an outage or damage to utility infrastructure.
Comfort Manyame (firstname.lastname@example.org) is the GIS manager for Mid-South Synergy Electric Coop in Texas. He earned his Ph.D. degree from Texas A&M University, College Station. He also is the technology editor for The GIS Professional, a URISA publication. His work on utility GIS, vegetation management and lightning strike studies have been widely published, including in ESRI's GIS Best Practices for Municipalities, Cooperatives and Rural Electric Utilities.
Key Soil Names, IDs and Characteristics
|Mapping unit name||MUID||AWC||Clay (%)||Organic matter (%)||Permeability inch/hr (mm/hr)||Dead tree outages||Dead trees cut||Model weight|
Soil mapping unit influence on dead tree-related outages.
GIS-Weighted Overlay Model
The initial step was to convert the soil type and vegetation cover shape files into raster data format to facilitate their manipulation in spatial analyst. Once converted to raster format, each of the soil and vegetation cover types were assigned a grid value and used in the analyses that followed.
The resultant raster output, while informative, still needs to be subjected to scrutiny by applying some ground truth. Maps showing the vegetation cover types, soil mapping units, dead trees and dead tree-related outages were produced.
Being able to show the spatial distribution of soil types and vegetation cover types across Mid-South Synergy Cooperative's service territory helps to highlight the need for site-specific vegetation management. Pines are the dominant vegetation cover type in Mid-South Synergy's territory, which may help to explain why 75% of dead trees cut in 2011 were pines.
Clearion | www.clearion.com
Mid-South Synergy Electric | www.midsouthsynergy.com