November 2025 · Field observation
Black River was assessed before anyone from CASSM traveled there. Four sources of different kinds were read against one another over roughly twenty-four hours on November 2, 2025, five days after landfall: federal aerial imagery, a citizen video shot on the ground by Jamaican content creator YowNoel, the recollection of a practitioner who knew the town, and a weather forecast read for a purpose it was not written for. None of the four would have been sufficient. Each covered a blind spot in at least one of the others, and the account below is as much about what each one could not show as about what it did.
What was available, and what each source could not reach
The assessment ran in a WhatsApp group comprising Dr. Franz Collins, Mrs. Gwen Miles, and Dr. Donovan Miles, with Dr. Kristoff Frank contributing through Dr. Collins. That is worth stating plainly rather than dressing up as a process. The coordination happened in an ordinary messaging application, in real time, between people who were also doing other things.
Federal aerial imagery. The National Geodetic Survey published emergency response imagery of the affected areas, georeferenced and viewable at building level. It resolved structural questions that nothing else could: which roofs were gone, which walls were standing, what a given compound looked like from above. What it could not show was ground condition. An aerial photograph does not distinguish firm ground from saturated ground, does not reveal standing water beneath tree canopy, and says nothing about whether a road surface will carry a loaded bus.
Citizen video. A vlog posted on November 1 by YowNoel documented conditions across the district at ground level and included interviews with residents. It supplied precisely what the aerial imagery could not: road surfaces, the state of routes into the area, the absence of power across the district, and an account of circumstances in residents' own words. Its limitation was that it was shot for the creator's purposes rather than CASSM's. The framing was not directable, and there was no way to know in advance whether something a walkthrough would have caught had been left out of shot.
Local knowledge. Dr. Kristoff Frank knew Black River. When CASSM circulated map material, he corrected the reading of it, identifying which side of a particular road had flooded relative to the prep school. That is the class of correction only familiarity produces; no image or forecast would have surfaced it. Its limitation is the general one affecting local knowledge after a disaster, that it describes a place as it was, and a hurricane is precisely the event that makes prior knowledge unreliable. Dr. Frank's assessment held. The category it belongs to does not always.
Weather forecasting. A five-day outlook for Black River was read not for whether it would rain but for whether the ground would be firm enough to take a tent configuration and vehicle movement, which identified the days offering the best drying conditions and those to avoid. This was the only forward-looking source of the four. Its limitation is inherent: a forecast is a probability, not an observation, and it degrades with distance from the day it is issued.
The four are not interchangeable and not ranked. Aerial imagery answered structural questions, the vlog answered access questions, local knowledge corrected interpretation, and the forecast answered a question about the future that none of the other three could address at all.
Assessing sites before traveling
The site inspection took place on November 6, 2025 and the team deployed on the 8th, nine and eleven days after landfall. Candidate locations were assessed against CASSM's operational requirements before any of that, and several were ruled out on what the combined sources showed.
Independence Park was the most significant of these. As the principal open space in Black River it was the natural site for a field clinic, but the ground was waterlogged while flood water remained, which would have made field work difficult and vehicle movement worse. That judgment did not rest on any single source. The standing water was visible on the ground; the forecast indicated how quickly conditions could be expected to firm; and local knowledge established which areas drained and which did not. Other locations were excluded because buildings had sustained structural damage and could not be treated as safe, the courthouse among them, or because the available footprint appeared too small for the tent configuration together with parking for five transport buses.
Those are the criteria that would have been applied on a walkthrough. The tent spacing and vehicle footprint were already documented, having been established during the resourcing work described in Sizing a field clinic under funding uncertainty, which meant the material was read against a known spatial requirement rather than a general impression of suitability.
The saving is operational rather than analytical. A tent configuration and five buses arriving at a site that cannot take them costs a deployment day in the acute phase, with clinical staff present and unable to work.
What the site visit confirmed
The pre-deployment visit on November 6, conducted by Dr. Franz Collins, Mrs. Gwen Miles, Ms. Christina Gordon of Jamaica AIDS Support for Life, and Dr. Donovan Miles, corroborated the remote assessment on each point that mattered, namely the extent of devastation, roads that were impassable or close to it, and the lighting constraint that ultimately set the operating window.
The visit served a second purpose the remote assessment could not. Both organizations were sending staff, and the assessment was made jointly rather than by CASSM alone. Ms. Gordon was assessing the site for a JASL team as much as for the deployment as a whole. That matters because agreement reached on the ground, between the people who will answer for their own staff, is a different thing from agreement reached over shared images. It is also faster. Questions that would have taken days of correspondence to resolve were settled in the time it took to walk the site.
That is the finding, and it is a limited one. One deployment in which cross-read sources proved accurate does not establish that the method is reliable in general, and CASSM verified on the ground before committing rather than relying solely on the remote assessment.
The elimination decisions carry a further caveat. The sites ruled out were not subsequently reassessed, so those judgments remain unverified. It is possible that a location excluded on the combined evidence would have served, and CASSM has no way of establishing that now.
Why published reporting did not serve
Press coverage was available throughout and was the least usable of anything CASSM consulted for a specific question. The St Elizabeth Parish Court in Black River was described in a Gleaner photograph caption on December 15 as damaged (UDC mobilises for Black River clean-up, 2025) and, in a Gleaner first-person account in January, as having caved in (Rodriguez-Moodie, 2026). Two accounts six weeks apart, with the more severe description arriving later. That is not a criticism of the reporting, which was written for general readers rather than for anyone trying to establish whether a particular structure could house a clinic. It is an observation about fit.
The contrast with the aerial imagery is instructive. The condition of that building could be determined from a georeferenced photograph within days by anyone who knew the imagery existed. The difficulty was never that information about Black River was unavailable; a great deal was published, and YowNoel's video was itself among the earliest accounts to report how much of the town had been destroyed. The difficulty was that no single source answered the specific operational questions CASSM had to settle, which is why they were read against one another rather than relied on individually.
Planning input offered from Kingston carried a different problem. Suggestions as to suitable locations drew on knowledge of Black River as it had been before the hurricane, which was confident, locally grounded, and in several respects no longer true. That is not an absence of information but information overtaken by events, and it is harder to identify as unreliable than silence is. It is also the same category as Dr. Frank's contribution, which held. Local knowledge is not uniformly reliable or unreliable after a disaster; whether it holds depends on what changed, and that has to be checked rather than assumed.
Why we made contact
CASSM contacted YowNoel on the morning of November 2, hours after finding the video and four days before the site visit, and again on the 5th. The purpose was to acknowledge the quality of the work, to ask whether he could advise on conditions at other sites, and to discuss working together on future deployments should circumstances allow.
The distinction matters. Treating citizen media as a passive feed extracts information from someone's labor without acknowledgement. Treating the creator as a source establishes a relationship that can be returned to, and recognizes that local knowledge held by residents is a capability rather than raw material. The footage discussed here is published at this link.
The wider gap
Information collection and sharing have been identified as a central problem in disaster medicine, with formal platforms developed specifically to connect agency information systems during response (Kondo et al., 2019). What CASSM assembled was not that. It was four unconnected sources, found separately, held in a messaging application, and cross-read by whoever happened to be in the group. The aerial imagery was published and freely available, and CASSM found it because someone thought to look. The video was discoverable only by chance. Nothing linked any of them, and nothing would have told a smaller or less networked organization that any of it existed.
The question this raises
The transferable practice is not the specific four sources. It is the discipline of asking, for each one, what it structurally cannot show, and then looking for something that covers that gap. An aerial image cannot report ground condition. Ground footage cannot answer what it was not pointed at. Local knowledge describes a place as it was. A forecast is not an observation. Once those limits are stated, the gaps become searchable.
Whether a repository could be assembled that made such sources discoverable together in the Jamaican context, and what would make it usable rather than merely large, is a question CASSM cannot answer from one deployment. It would require aggregating material across platforms, establishing a basis for assessing source reliability, and arranging with creators in a collaborative rather than extractive way. Each of those is a design problem in its own right.
What this deployment establishes is narrower: that a usable remote assessment was possible five days after landfall, from sources that were public or a message away, and that no single one of them would have been enough.
Acknowledgements
YowNoel documented conditions across St. Elizabeth in the days after landfall and responded to CASSM's approach. Dr. Kristoff Frank provided the local assessment that corrected CASSM's reading of the map material, and Dr. Franz Collins relayed it. Mrs. Gwen Miles participated in the pre-deployment assessment throughout. Ms. Christina Gordon of Jamaica AIDS Support for Life took part in the site visit and was present on the day of deployment.
References
Kondo, Y., Ichikawa, M., Kondo, H., Koido, Y., & Otomo, Y. (2019). Current disaster medicine in Japan and the change brought by information sharing. Journal of Disaster Research, 14(2), 292–302. https://doi.org/10.20965/jdr.2019.p0292
National Geodetic Survey. (2025). Hurricane Melissa emergency response imagery [Data set]. National Oceanic and Atmospheric Administration. https://storms.ngs.noaa.gov/storms/melissa/
Rodriguez-Moodie, T. (2026, January 4). After Hurricane Melissa, Black River waits. The Gleaner. https://jamaica-gleaner.com/article/focus/20260104/theresa-rodriguez-moodie-after-hurricane-melissa-black-river-waits
UDC mobilises for Black River clean-up following Holness order. (2025, December 15). The Gleaner. https://jamaica-gleaner.com/article/news/20251215/udc-mobilises-black-river-clean-following-holness-order
YowNoel. (2025, November 1). Black River destroyed: Hurricane Melissa leaves St. Elizabeth in darkness [Video]. YouTube. https://www.youtube.com/watch?v=cRfbCgTgfRk