SUMMARY
Maui Island, in the central Pacific Ocean, is prone to wildfires, which are often exacerbated by dry and windy conditions. The frequency of these wildfires has increased in recent years, most likely due to a combination of climate change, droughts, and human activities.
The island’s vulnerability to wildfires is increased by overgrown vegetation and non-native invasive plant species, such as guinea, molasses, and buffel grass. These factors contribute to the intensity and rapid spread of wildfires.
The state is prone to wildfires due to its dry climate and steep terrain. In recent years, there have been a number of large wildfires in Hawaii, including the Big Island fires in 2018 and the Puna fires in 2019.
On August 8th, 2023, a brush fire erupted in Kula, approximately 35 miles (56 km) from Lahaina. Reportedly, the Hawaii Electric Company (HECO) said it “appears to have been caused by power lines that fell in high winds.” The fire was fueled by the dry conditions of the Hawaiian Islands, for which the US National Weather Service issued an alert. The wildfire moved rapidly towards the city of Lahaina under the strong wind conditions generated by a high-pressure area to the north of the island, and Hurricane Dora, which was crossing the Pacific Ocean to the south.
The fire caused extensive damage and is considered the deadliest wildfire in Hawaii’s history.
LAHAINA OVERVIEW
Lahaina is a popular tourist destination on the western coast of Maui, Hawaii, known for its beaches, whale watching, and surfing. The town is known for its historic buildings, including the Lahaina Courthouse, the Old Lahaina Jail, and the Baldwin House and also a home to a number of museums, art galleries, and shops.
Skytek focused on the Lahaina town to monitor the wildfire event, the overall exposure and an assessment of the affected area.
LAHAINA WILDFIRES OF AUGUST 08TH 2023 OVERVIEW
Skytek investigated the spread of the wildfire using the thermal signature on the island captured by a NASA satellite with thermal capture capabilities which is represented for visualisation in Figure 1.
Pre and post high-resolution satellite imagery was acquired by Skytek to detect, highlight and quantify the changes associated with the wildfire damage.
In addition, Skytek ran its proprietary change detection algorithm to analyse the impact of the wildfires on the Lahaina city and produced a visual layer of detected damage which is displayed superimposed over the high resolution imagery captured for August 15th 2023., displayed in Figure 2 below.
Figure 2 – Planet Labs satellite imagery of Lahaina city on February 02nd, 2023
Figure 3 – Digital Globe satellite imagery of Lahaina city, damage overlay on August 15th, 2023
Skytek provides two views to support damage assessment
- Macro level: To detect damage over large areas, e.g. city region
- Micro Level: Each property included in the change detection algorithm has a support pre/post image
MACRO LEVEL
The damage detection algorithm identifies the boundaries of each property in a set of pre- and post-wildfire images and produces a damage layer highlighting the identified clusters of damaged properties for the entire area of interest
The damage from the wildfire is concentrated in Lahaina’s old city, where many of the properties are built of wood and have wooden features. A cluster of properties that have escaped the fire is around the Kahoma village, without many trees around which could have propagated the fire.
Skytek’s damage detection algorithm has analysed the 1,199 properties captured in Figure 2 and aggregated the damage detected in the table displayed in Figure 4.
Damage Category | Damage Range | Number of Properties | Property Damage % | |
 | RED | 76%-100% | 773 | 64.47% |
 | ORANGE | 26%-75% | 84 | 7.01% |
 | GREEN | 0%-25% | 342 | 28.52% |
Figure 4 – Skytek damage assessment overview for Lahaina town, for August 10th 2023
MICRO LEVEL: Individual Properties
Each of the 1,199 analysed properties included above has its own individual pre/post imagery and damage assessment.
The damage score is listed granularly, with each property identified by its coordinates and address. The results of the damage detection model can be visualised in a table, as shown for a sample list of properties in Figure 5 below:
ID | Address | Lat. | Long. | Damage % | Damage Category | |
| | 1033 | 611 Front St., Lahaina, HI 96761, USA | 20.8702º N | -156.6769º W | 81.18% | Red |
| | 1034 | 127 Lahainaluna Rd., Lahaina, HI 76761, USA | 29.4393º N | -083.2868º W | 100.00% | Red |
Figure 5 – Sample table of property damage analyse, Lahaina, Maui
Skytek has completed detailed Machine Learning assessments for all affected properties. Two sample reports are presented below.
Upon request, Skytek can provide detailed reports for all properties .
ID | Street Address | City | State | Zip | County | Report |
1033 | Lahaina | HI | 96761 | Maui | ||
1034 | Lahaina | HI | 96761 | Maui |
