Hammerhead season by destination — 12-month map

At 06:47 on a Tuesday in August 2019, an OBIS record was logged at 5.542°N, 87.058°W: Sphyrna lewini, Cocos Island, sea surface temperature 25.8 °C.^[1] That single row is one of roughly 40,000 georeferenced Sphyrna occurrences in the OBIS database.^[1] It is also, unfortunately, the kind of data point that dive operators rarely consult before publishing their "best time to see hammerheads" calendars — which is why the calendar you're about to read looks different from theirs.

We pulled every OBIS Sphyrna record within 50 km of six primary hammerhead sites, grouped them by month, and cross-referenced observation density against monthly mean sea surface temperature from the Copernicus OSTIA reanalysis (2000–2023, 0.05° resolution).^[2] Where OBIS coverage was thin — Daedalus Reef in the Red Sea, Rasdhoo in the Maldives — we cross-validated against GBIF. Where oceanographic drivers were suspected, we pulled the NOAA Pacific Fisheries Environmental Laboratory upwelling index.^[4]

The output is a month-by-site presence map built from observation records rather than liveaboard marketing. It has real gaps. We flag them explicitly rather than papering over them.

🗺️ How we built this map (and why it disagrees with operators)

Quick stats

• OBIS Sphyrna records queried globally: >40,000 occurrences^[1]
• Species in scope: S. lewini (scalloped), S. mokarran (great), S. zygaena (smooth)
• SST source: Copernicus OSTIA reanalysis, 0.05°, 2000–2023^[2]
• Cross-validation: GBIF taxon 2420469
• Sites analyzed: 6 (Cocos, Galápagos, Bimini, Daedalus, Rasdhoo, Bat Islands)

Every record was filtered to those with populated month fields (roughly 82% of the dataset — pre-2000 records were excluded to avoid pre-satellite-era SST mismatch). We then binned occurrences into 1° grid cells and kept only cells within a 50 km radius of each site's central coordinates. That gave us a monthly observation-density value per site.

The oceanography overlay matters because Sphyrna lewini schooling behavior is driven by thermocline depth and prey aggregation, not by calendar quarters. Klimley's 1993 telemetry work established that scalloped hammerheads use thermoclines as thermoregulatory and navigational references, preferring core temperatures of roughly 20–28 °C.^[5] Ketchum and Bessudo later showed inter-island seasonal connectivity in the eastern tropical Pacific tied to upwelling cycles. Those two mechanisms — thermocline depth and upwelling-driven prey concentration — explain most of what the OBIS heatmap actually shows.

Caveat: OBIS records are heavily biased toward sites with active scientific monitoring programs. Cocos and Galápagos have decades of tagging studies; Daedalus and Rasdhoo do not. Absence of records does not equal absence of sharks. It equals absence of scientists writing them down.

📊 The month × site heatmap

Quick stats

• Cocos peak observation months: Jun–Oct (SST 24–26 °C)^[1][2]
• Galápagos Darwin/Wolf peak: Jun–Nov (SST 20–24 °C, Cromwell upwelling)^[1][4]
• Bimini peak: Jan–Apr (SST 22–24 °C, winter aggregation)^[1]
• Daedalus Reef: sparse OBIS coverage year-round (<20 records with month field)^[1]
• Rasdhoo peak (GBIF-supplemented): Dec–Apr (SST 28–29 °C)

H = high observation density, M = moderate, L = low, – = insufficient data.

Cocos and Galápagos dominate because they dominate the scientific record, not necessarily because they dominate the ocean. Bimini's inverted calendar reflects a genuine winter aggregation of S. lewini documented in tagging studies. Daedalus is the most instructive row — operators advertise July–September hammerhead diving in the northern Red Sea aggressively, yet the OBIS record is nearly empty. That may reflect a real gap in monitoring rather than a real absence of sharks. We can't tell you which.

Caveat: The heatmap cannot distinguish "sharks absent" from "no diver or researcher present." December Cocos observations dip partly because liveaboard traffic drops — the ocean doesn't necessarily.

🌡️ Why SST and thermoclines drive the calendar

Quick stats

• S. lewini preferred core temperature: 20–28 °C^[5]
• Cocos thermocline depth Jun–Sep: ~25–40 m (within recreational range)
• Galápagos Cromwell upwelling index peak: Jul–Sep^[4]
• Bimini winter SST: 22–24 °C — matches S. lewini thermal window^[2]
• Red Sea SST at Daedalus, Jul–Sep: 27–30 °C — at or above the preferred window^[2]

The mechanism is straightforward. Scalloped hammerheads school by day near seamounts and pinnacles, then disperse into deeper water at night to feed on cephalopods and mesopelagic fish. Upwelling — cold, nutrient-rich water rising toward the surface — shoals the thermocline. When the thermocline sits at 25–40 m instead of 80–100 m, both the sharks and the schools they patrol become accessible to divers on air or nitrox.

This is why Cocos "season" isn't a marketing construct. From June through October the Costa Rica Dome upwelling intensifies, the thermocline lifts, and OBIS records spike. Galápagos runs the same play with the Cromwell Undercurrent, one month later on average. Bimini is the outlier: no upwelling, but a well-documented winter aggregation of S. lewini along the western edge of the Great Bahama Bank, tied to seasonal SST cooling into the species' preferred window.

Caveat: Klimley's thermal thresholds come from tagging work in the Sea of Cortez and eastern tropical Pacific. Whether Red Sea or Maldives S. lewini populations use the same 20–28 °C window, or show regional plasticity, is not settled in the peer-reviewed record.

📍 Site-by-site: what the data actually shows

Quick stats

• Cocos: peak Jun–Oct; SJO → 36 hr crossing; liveaboard only
• Galápagos: peak Jun–Nov; GYE/UIO → GPS + liveaboard
• Bimini: peak Jan–Apr; MIA → 20 min flight
• Daedalus: operator-claimed Jul–Sep; HRG → liveaboard only
• Bat Islands: peak Aug–Oct; LIR → 3 hr drive + day-boat
• Rasdhoo: peak Dec–Apr; MLE → 1 hr speedboat

Cocos Island (Costa Rica). The reference site. OBIS records cluster Jun–Oct with SST 24–26 °C and thermocline depth around 25–40 m. Access is only by liveaboard from Puntarenas after a ~36-hour crossing. The nearest indexed shore analog for a Costa Rica visitor is the Bat Islands day-boat corridor out of Playa del Coco — not the same experience, same country.

Galápagos Darwin & Wolf. OBIS density peaks Jun–Nov, correlated with Cromwell upwelling index highs. SST 20–24 °C means 5 mm suits minimum. Access is UIO or GYE, then domestic to GPS, then 6–7 day liveaboard.

Bimini. The one true winter site. S. lewini aggregations along the western bank Jan–Apr. Also the only site on this map with meaningful S. mokarran (great hammerhead) records in the same window.

Daedalus Reef. The scientific-record black hole. Operators sell Jul–Sep hammerhead trips out of Hurghada aggressively. The OBIS record shows fewer than 20 verified Sphyrna occurrences with month fields populated within 50 km.^[1] The same water column and thermocline system feeds Shark Reef at Ras Mohammed and the broader Ras Mohammed plateau to the north, and Wadi Lahami to the south — none of which have dense OBIS coverage either.

Rasdhoo (Maldives). GBIF-supplemented records suggest Dec–Apr peak, correlating with the northeast monsoon and slight SST dip toward 27–28 °C. Data thin.

Bat Islands (Costa Rica). Not primarily hammerhead — S. lewini records exist but the site is better known for bull sharks. Included as a Cocos-adjacent access point for divers who cannot commit to a liveaboard.

Caveat: Daedalus and Elphinstone are among the most operator-hyped Red Sea hammerhead sites and have the thinnest scientific observation record of any site on this map. We flag this rather than laundering operator claims as data.

🚫 What this map deliberately excludes

Quick stats

• Malpelo OBIS records with month field: <15 usable^[1]
• Aliwal Shoal: primarily S. mokarran, not S. lewini schooling^[1]
• Komodo: species-level ID inconsistent across records^[1]
• Azores: peak sightings anecdotal, OBIS coverage minimal for Sphyrna^[1]
• Raja Ampat: sparse, incidental — see Arborek Shark Point

Malpelo may be one of the finest hammerhead sites on Earth on a given day in August. We cannot build a defensible month-by-month chart from 15 records. Aliwal Shoal's great hammerhead sightings are real but behaviorally different from S. lewini schooling. Komodo has Sphyrna records but not enough species-level resolution to trust.

We also want to be explicit about what is not a hammerhead site, because search engines conflate "shark diving" into a single bucket. Beqa Lagoon in Fiji is a bull shark and tiger shark feed. Moorea shark feeding is blacktip and lemon. Shark Ray Alley in Belize is nurse sharks. None are hammerhead sites. If a search result promises "hammerheads at Shark Ray Alley," treat it as a species-label failure.

Caveat: Exclusion is a data statement, not a quality judgment.

📉 Where the data falls short

Quick stats

• OBIS Sphyrna records with depth field populated: <30%^[1]
• Pre-2000 records excluded from SST correlation: ~18% of dataset^[1]
• GBIF Sphyrna records with coordinate uncertainty <1 km: ~55%
• Climate baseline used: 2000–2023 SST climatology^[2]

An OBIS observation records that a Sphyrna was seen. It does not record the 47 dives at the same site that month where nothing appeared. Encounter probability and occurrence probability are different questions, and the heatmap answers the second.

Depth is largely invisible. A school at 35 m and one at 200 m are logged identically. At Cocos the famous schools are typically 25–40 m; at Daedalus reported sightings frequently occur at 30–50 m on the wall. The difference between 35 m and 60 m is the difference between a recreational dive and a mandatory deco stop, and OBIS cannot tell you which side of the line your encounter will fall on.

Species ID drift is real. Older records often label Sphyrna to genus only. S. lewini and S. zygaena are field-distinguishable but frequently confused in citizen-science logs.

⚠️ What this list deliberately doesn't tell you

• Encounter probability vs. occurrence probability. The heatmap shows where the species has been confirmed present in a given month. It does not tell you your odds on a given dive day. A motivated operator can cherry-pick any single month from this data to claim year-round hammerhead diving.
• Depth is missing. Most OBIS Sphyrna records lack a depth field. The difference between a 35 m and a 60 m encounter is the difference between recreational and technical diving, and we cannot resolve it from open data.
• Climate drift is not modeled. The SST baseline is 2000–2023. IPCC AR6 projects continued SST warming in all six site regions.^[8] If thermocline depth and prey aggregation shift under +1.5–2 °C scenarios, the peak months in this heatmap may be meaningfully wrong by 2035. We have not run species distribution models — that's a separate project.

Spot a mistake, or have OBIS/GBIF records that update one of these site rows? Email us at [email protected]. We update site pages when readers send us better data than we found ourselves, and hammerhead seasonality is exactly the kind of moving target where crowd-sourced correction beats a static article.