How Advanced Technologies Helped in the Aftermath of Hurricane Otis: A Case Study from Guerrero, Mexico
- Matthew Cua | Innovation and IT Director
- Apr 24
- 7 min read
Updated: Apr 27
On 25 October 2023, the state of Guerrero, Mexico, was devastated by Hurricane Otis, surpassing Hurricane Patricia as the strongest Pacific hurricane on record. The hurricane was one of the most consequential forecast model misses in recent history, as its explosive intensification caught the city of Acapulco, home to approximately 658,000 people, completely off guard.
The hurricane's winds, with gusts reaching up to 329.76 km/h—making it the seventh-highest anemometer-measured wind gust recorded worldwide—destroyed critical infrastructure and communication networks. Labeled a "nightmare scenario," the widespread destruction left almost no part of the city untouched. This posed significant challenges for local authorities and response teams, who struggled to coordinate recovery efforts due to severely disrupted telecommunications systems.
Restoring Connectivity : Starlink & Intelsat Antennas
Upon arrival in Acapulco City, Mexico via the staging location at Chilpancingo City, Mexico the immediate priority was to restore connectivity for various stakeholders. In collaboration with the Guerrero Civil Protection Secretariat, Help.NGO deployed advanced communication technologies, including Starlink satellite antennas and Intelsat FL60P-E Starwin satellite antennas. These technologies significantly improved communication capabilities, facilitating real-time decision-making and enabling rescue teams to access previously unreachable areas.

With traditional networks down, Starlink and Intelsat satellite antennas became critical for reestablishing connectivity. Both satellite systems provided continuous links between field teams and operations centers, ensuring effective data transmission despite disrupted traditional infrastructure. Starlink antennas, known for rapid deployment and reliable connections, provided essential coverage in remote areas, while Intelsat’s established network ensured comprehensive connectivity across disaster zones. This dual-layered communication infrastructure was pivotal in coordinating rescue operations and distributing humanitarian aid efficiently.

Drones for Mapping and Rescue Operations

In addition to satellite connectivity, Help.NGO deployed several drones to map the city, rapidly assess damage, locate trapped individuals, and aid planning for recovery and reconstruction efforts.
The drones used included:
DJI Mavic 3 Enterprise (M3E): A versatile drone for mapping, search and rescue, and documentation.
DJI Mavic Air 2S: Cost-effective drone suitable for mapping, search and rescue, and documentation. Primarily used for partner training.
DJI Mavic Mini 2/SE: Primary training drone for novice operators, used especially in high-risk scenarios.
EbeeX: AgEagle’s large-scale mapping drone, capable of approximately 90-minute flights covering extensive areas. It was primarily used for data uploaded to OpenAerialMap, such as the location in Jardin
Ultimately, Help.NGO collected over 142GB of data, mapping 51 square kilometers, with most data uploaded to OpenAerialMap.
Snowball Edge from AWS: Efficient Data Storage and Processing
Despite satellite innovations, cloud connectivity remained limited. Given the massive volume of drone-generated data, uploading directly to the cloud was impractical during emergency operations. Help.NGO utilized AWS Snowball Edge, specifically the Compute Optimized GPU model featuring 39.5TB HDD and 7.68TB SSD, preloaded with WebODM on an Amazon EC2 instance. This allowed on-site processing of drone images, with only processed data uploaded to optimize limited field bandwidth.
WebODM’s user-friendly interface on AWS Snowball Edge facilitated data sharing among stakeholders within the same network, significantly reducing processing time and enhancing decision-making efficiency.

OpenDroneMap for Image Processing and Mapping
The OpenDroneMap ecosystem is an open-source drone mapping solution. Most organizations, especially non-profit organizations and organizations in developing nations, face critical challenges with their current drone processing and visualization workflows as bundled drone processing software and licenses impose significant budgetary strain. OpenDroneMap not only provides an open-source solution but also offers several extensions such as WebODM that creates a user-friendly web-based application to upload, process, configure and visualize drone data.
Help.NGO deployed WebODM on AWS Amazon EC2 both in the cloud and on an AWS Snowball Edge device. We deployed WebODM on a C5.4xlarge Amazon EC2 instance with Ubuntu, which was compatible with AWS Snowball Edge and was more than adequate to process drone image projects of less than 999 images per project. C5.2xlarge can also be used, but we were only able to process no more than 500 images per project, though it can easily load pre-processed projects.
Because of OpenDroneMap's community and architecture, once processed, the entire project can be exported and loaded into another WebODM instance. For this operation, maps processed in the AWS Snowball Edge were uploaded into the AWS Cloud-based WebODM, and those processed in the AWS Cloud-based WebODM were downloaded and loaded into the WebODM on AWS Snowball, allowing flexibility of operations and decision making.
Having this flexibility reduced the back and forth of data transfers and sped up decision making. A drone field team can fly the drone for 90 minutes, process the data in the AWS Snowball Edge in less than 30 minutes, upload the processed data via Starlink into the AWS Cloud-based WebODM in less than 1 hour, and then decision makers 4 hours drive away can open their low-cost tablet and be able to view and manipulate the drone data to make decisions.
With WebODM on AWS Cloud and AWS Snowball Edge, Help.NGO was able to turn drone-captured images into detailed maps, orthomosaics, and 3D models that accurately reflected the extent of the damage in almost real-time, providing precise geospatial information that helped response teams and decision makers make informed decisions and allocate resources more effectively.
Upon completion of the project, the raw drone images stored in the AWS Snowball Edge were automatically uploaded into an Amazon S3 Bucket, reducing the need to manually upload raw data.

Collaborative Mapping with HOTOSM
Collaboration with the Humanitarian OpenStreetMap Team (HOT) was vital for improving response coordination. Help.NGO and HOTOSM have been collaborating for several years, mostly remotely but in this case also in the field. The HOT Latin America and the Caribbean (LAC) Hub was in the field together with Help.NGO to assess post-disaster damage using Help.NGO's and CENAPRED drone imagery, as well as satellite imagery provided by partners such as Copernicus and other sources, to collect geospatial identification of destroyed buildings and other geospatial characteristics such as unmapped roads necessary for the actions of humanitarian organizations and the government itself. HOTOSM then makes this data available on OpenStreetMap for any one to use.
HOTOSM led a humanitarian mapping activation, collaborating with universities, local authorities, international organizations and volunteers to map over 22,000 damaged structures and key public facilities and other kinds of operational data.
Aside from using OpenAerialMap and Tasking Manager, HOTOSM also uses KoboToolbox. By using these tools, the HOTOSM team is able to support early response and recovery decisions -- such as identifying shelters and supply distribution points.

Find out more about the work of HOT in Acapulco in the following blog posts:

Collaboration with the Guerrero Civil Protection Secretariat
The partnership between Help.NGO and the Guerrero Civil Protection Secretariat was key to ensuring a swift and effective response. By working closely together, both organizations overcame the challenges posed by destroyed infrastructure and coordinated efforts to provide assistance.
Restoring Connectivity and Coordinating Response: The deployment of Starlink and Intelsat satellite antennas enabled constant communication between field teams and command centers. This connectivity allowed local authorities to coordinate rescue operations, distribute supplies, and manage resources efficiently.
Help.NGO team members deploying Starlink with World Central Kitchen Using Drones for Assessment and Rescue: The various drones were instrumental in identifying trapped individuals, disconnected communities, and creating detailed maps of affected areas for post-disaster recovery and rehabilitation purposes.
Help.NGO Operations Officer Milo Lara, coordinating with local rescuers in debris clearing operation. Training for Civil Protection Agencies: In addition to deploying technology, Help.NGO also conducted training for all agencies within the Guerrero Civil Protection System, with more than 42 responders trained over a 5-day period, including 2 days in the field. The training focused on drone operations, cloud computing, and geospatial image processing. Sessions included both theoretical knowledge and hands-on experience with drones for search and mapping missions, as well as handling large volumes of data. This training aimed to equip the agencies with the tools and knowledge needed to incorporate these technologies into future emergency responses.
Training of Civil Protection Agencies in Chilpancingo City, Mexico
The Impact of the Deployment: A More Efficient and Coordinated Response
The deployment of Help.NGO's technologies in Guerrero, Mexico during the aftermath of Hurricane Otis significantly improved the operational capacity of local authorities and provided a more efficient, coordinated response.
By providing both technological solutions and training for local agencies, Help.NGO helped ensure that future responses to disasters like Hurricane Otis will be more agile and better coordinated, ultimately saving lives and speeding up recovery efforts.
Partner Collaboration: A Key to Success
The success of this project is a testament to the strength of the partnerships of Help.NGO
We would like to thank the following private partners:
AWS Disaster Response Team and the various AWS volunteers for their guidance in architecture design and working with us in the field as well as financially supporting this operation.
MAD Services for providing subject matter expertise (SME) in developing and deploying the various Edge and UAS processing solutions.
Starlink for providing the LEO connectivity as well as technical support
Intelsat for providing the MEO and GEO connectivity as well as technical support
Their generosity was integral to the project’s success.

Further Reading:
Written by :
Matthew Aligaen Cua, Innovation Director at HELP.NGO and an Adjunct Faculty at the Asian Institute of Management (AIM). With a background in integrating technology into humanitarian efforts, Matthew has led numerous missions worldwide, focusing on enhancing disaster response through innovative solutions.
Andrés Guevara, Deputy Director of Operations at Help.NGO. He has more than a decade in experience in leading teams and organizations in emergency management, firefighting. first aid and disaster risk reduction activities worldwide. Mr. Guevera, not only holds an Emergency Trauma Bag Trainer certificate he also holds a UAS license and has more than 5,000 hours of flight experience teaching practitioners worldwide. He is passionate about integrating technology in humanitarian efforts while looking tacticool with it.
Matthew Silver, U.S.-based Research Consultant at Help.NGO. He supports the organization’s global operations by leading critical research initiatives, overseeing communications quality, and advancing strategic development efforts across multiple domains. Matthew holds a B.S. in Sociocultural Anthropology and Political Science from Colgate University and is passionate about applying rigorous analysis to humanitarian efforts, bridging academic insight with real-world impact.
Contact Information:
Email: mcua@aim.edu, matt@help.ngo
For more information on HELP.NGO’s work, please visit www.help.ngo.
Comments