Methods: what the Ocean Count numbers mean

Every results sheet we use, 2013 to 2026, is titled “Preliminary Data”, and every number from those sheets carries that label. The sanctuary collects and publishes the observations. We compile them and are not part of NOAA or the Ocean Count program.

What the island average measures

On the 2026 sheets, each site’s count is split into nine 15-minute periods, one starting every half hour from 8:00 am to 12:00 pm, with adults and calves logged separately. Every sheet we use prints one figure per island, labeled “Average Number of Whales per 15min Count Period” (the 2026 sheets write “15 min.”), and none says how that average is calculated.

The 2013, 2014 and 2026 sheets also print every site’s count for every period, so we recalculated their 29 island figures. Of those, 28 match the average of adults plus calves, with periods marked N/A left out and rounded the way the sheet prints it: 27 when every site-period counts equally, and one more when each site is averaged first. Counting adults only, 11 of the 29 match.

Of the other two, Oʻahu in January 2014 (printed 2) matches only if calves are left out; Hawaiʻi in March 2014 (printed 2) matches if you average each site over the periods it counted, then average the sites (rain ended several of its sites early that day). So the average of adults plus calves per site-period is our best reading of the sheets, not a definition the sanctuary has published.

The 2012 sheets print site tables but no island average; the same calculation on them reproduces all 9 averages the 2012 press releases state.

For example, on March 28, 2026, 10 Kauaʻi sites logged 90 site-periods between them. Their average is 1.88 whales per site-period, and the sheet prints 2. Sites where volunteers saw nothing still count toward the average, so adding sites with few sightings lowers an island’s figure even when volunteers at the other sites see as many whales as before.

The average is built from whales seen from shore. It does not estimate how many whales are in Hawaiian waters, and the same whale can be counted at two sites or in two periods. The sanctuary’s March 2026 release says its daily totals “may represent duplicate sightings of the same whale by different observers or at different time periods or different locations throughout the day.”

The 2007 to 2011 and 2015 to 2025 site tables

The 47 results sheets of 2007 to 2011 and 2015 to 2025 print each island’s site table as an image, 154 tables in all. We typed in every cell, 18,760 of them with a count. Where the press release gives an island’s day total, we checked the sum of our cells against it. Where it does not, we read the table a second time and settled every difference against an enlarged copy of the sheet.

The 2007 to 2011 tables were each read twice, independently. From 2008 the press releases give whole-number island averages, and we checked the average of our cells against them.

Averaged as adults plus calves, per site-period or site by site as for Hawaiʻi in March 2014, our cells reproduce the printed island average in 90 of the 99 tables outside the February 2015 sheet, whose averages print 0 for every island. They match the release’s day total in 24 of 33. Leaving out the four cells below raises the averages to 93; reading them at the corrected values raises the day totals to 28.

Read as the values in the table below, each of those four cells makes both the release’s day total and the printed average agree. We keep them as printed, flag them in the file and leave them out of our site averages. The rest of the disagreements are between the sheet and its release, or between a sheet’s average and its own table, and we keep the sheet as printed.

On January 28, 2023, the release total for Hawaiʻi Island equals the sheet’s total without Kapaʻa Beach Park. On March 29, 2025, the release counts one Molokaʻi site and the sheet adds a second, Sandy’s (printed Sandys). On February 24, 2024, the sheet’s file name says Updated, and the Hawaiʻi Island table’s extra 38 whales equal the row for Hualālai (Four Seasons Resort). Each count’s notes are in the file’s note column.

Preliminary and revised sheets

We copy each figure exactly as printed and link every row to its PDF. Of the 40 results sheets we use, 5 have “Updated” or “Final” in the file name, so an earlier version of those may have circulated. We use the version posted on oceancount.org or the sanctuary site when we collected them in September 2026, and for 2013 to 2015 the Internet Archive’s copies of the sheets from the sanctuary’s old site. If a sheet is revised again, our copy will differ until we rebuild the dataset.

Whole numbers and decimals

The sanctuary has printed the average with different precision over the years. From 2020 to 2023, 24 of 33 island figures carry a decimal, such as 2.3. In 2013, 2014, 2015, 2017, 2024, 2025 and 2026, every figure is a whole number. 2016, 2018 and 2019 mix the two: whole numbers for most figures, with decimals only for values of 1.5 or less.

A whole number can hide a difference of almost one whale, so a change from 2 to 2.3 between years may be rounding. Our CSV flags every value’s precision so you can filter or treat the two kinds separately.

On the 2013 and 2026 sheets every island figure matches half-up rounding of that average, so the 2026 printed 0 is an average under 0.5. On March 28, 2026, the sheet printed 0 for Hawaiʻi, but the island’s site table averages 0.49 whales per site-period: volunteers did see whales. The March 2025 sheet also printed 0 for Hawaiʻi; our typed copy of its site table, which is an image, averages 0.39.

Weather and site changes between counts

The number of Ocean Count sites that reported ranged from 29 to 62 in the 39 counts whose release gives it. Since 2024 the sheets say “Only active site locations for the year are included”, so the set of sites behind an island average can change between seasons. Our dataset does not adjust for this, or for the weather.

Wind and rain cut how far observers can see. In March 2026 the sanctuary reported “moderate to strong winds and the presence of white caps which significantly reduced visibility”. In January 2023, rain and storms caused poor visibility or ended the count early at several sites, and the Great Whale Count on Maui could report from only one of its 12 sites. Before comparing two counts, read the press release for each; the links are in the count-summary CSV.

The chart below shows, for every count from 2007 to 2026, how many listed sites on each island reported a site table, how many the sheet marked closed for weather, surf, a tsunami warning or a cancellation, and how many were listed but not used that month. Our site averages and the site heatmaps use only the sites that reported. In late March 2018, weather closed seven Oʻahu and six Kauaʻi sites; the site lists are shorter from 2021 on.

The COVID-19 years

The March 2020 count was canceled. From January 2021 to February 2022, the program ran without public volunteers: “each site is monitored by trained site leaders working individually or as a couple,” in the words of the January 2021 release. Fewer observers per site can mean fewer whales spotted, so treat those 5 counts, marked site leaders in the participant_role column, as a separate series. Volunteers returned for the March 2022 count.

Molokaʻi, Maui and the combined headcounts

Molokaʻi sites first appear on the results sheets in 2024, with 8 counts through March 2026. The 2013 to 2023 sheets cover Kauaʻi, Oʻahu and Hawaiʻi Island only. The Ocean Count sheets list no Maui sites. Pacific Whale Foundation runs its own Great Whale Count on Maui, and on Lānaʻi since 2024, on the same mornings and publishes the results on its community science page.

From 2019 the two programs have shared press releases, and 22 of the 23 headcounts since then cover both counts. The 21 headcounts from 2008 to 2015 cover the Ocean Count alone, so check the participants_include_maui_great_whale_count column before you compare headcounts across 2019. The releases also give day totals per island. Those totals add up sightings across sites and periods, so they can count the same whale more than once, and we publish them as given without summing them.

What is not in the dataset

Our site-period file starts in 2007, the first season with an archived results sheet; for 2002 to 2006 the 2010 report gives each site’s value per count instead. Behavior records are not on the results sheets, which state that “the behavioral data that is collected for the project is not part of this report.” For 2002 to 2010 the report gives each site’s breaches and slaps per whale (dives from 2006), which we typed in; they are on the download page.

For 2005 and 2006 our files hold the 2010 report’s per-site averages, and for 2007 the results sheets’ site tables. From the 2010 report’s whale counts we left out its island summaries and coastline comparisons; our behaviors page quotes its island summaries of breaches and slaps.

Island averages from the press releases, 2008 to 2012

The results sheets we found for 2008 to 2012 print no island average. The press releases for those counts state one in their text, so we publish them in a separate file, 51 rows from 15 counts. They come from a different source type than the sheets and every value is a whole number, so we keep them out of the results-sheet charts and statistics on our main data page and show them only in the chart that marks each figure’s source.

The 2012 figures match the site tables on that year’s sheets. The earlier sheets print their site figures as images; for 2008 to 2010, the 2010 report’s site averages are our check instead. The 2008 and 2009 releases also report Kahoʻolawe.

Where a release gives a note instead of a number, or next to one, the file keeps it in the release_note column: Kahoʻolawe, January 2008: “Did not participate in January count”; Kahoʻolawe, February 2009: “Cancelled due to inclement weather”; Hawaiʻi, February 2012: “affected by inclement weather”.

Seven releases from 2008 to 2011 also state an average per site for the whole state; it is in the statewide_avg_whales_per_site_per_15min column of the count-summary file. The January 31, 2009 figure of 8 does not fit the other figures for that count: the 2010 report’s 56 site averages for that count average 4.84, and the same release gives island averages of Kauaʻi 8, Oʻahu 4, Hawaiʻi 5, Kahoʻolawe 14. The statewide_note column flags it.

The sanctuary’s reports, 2002 to 2010

For 2002 to 2004, a report for each season, written under contract for the sanctuary by Daniela Maldini and dated after the season, states an island mean of whales per site per 15-minute period, with a standard error, in its text: 27 island figures from the 2002 report and the two that followed. The 2002 count has the same timing as the 2026 sheets: nine 15-minute scans between 8:00 am and 12:15 pm.

A later report, written in 2010 for the sanctuary by Ashton Mead Cullum, an Ernest F. Hollings Scholar intern, prints a table for each count site with its average whale sightings per 15-minute count period for every January, February and March from 2002 to 2010. The tables are images, so we typed them in by hand: 63 sites over 27 counts.

Every value in the report’s tables is a whole number. The report marks a closed site, a missing data sheet or unreliable data as NA, which our file leaves empty, and a 0 means no whales were seen.

To check our typing, we averaged the report’s site values for each island and count and compared them with the island averages in the 2008 to 2010 press releases. All 31 pairs differ by 1 or less, and only Kahoʻolawe in March 2009 (report 8, release 9) differs by 1.

We publish no island figures of our own for 2005 to 2007. For the report years, averaging the report’s whole-number site values would carry each value’s rounding into an island figure, and the 2007 sheets, whose site tables we typed in, print no island average to check a sum against. The report’s island summaries give yearly means for 2006 and 2007 (Kauaʻi 3.67 and 4.00, Oʻahu 2.67 and 3.33, Hawaiʻi 2.67 and 3.67) but no figure for a single count, so they are not in our files.

The report prints the same 2002 to 2005 values for Mokulēʻia Beach Park as for Kaʻena Point (West), probably a copy error; the note column flags those rows. Two Kauaʻi report sites, Mahaulepu Haula (#8) and Princeville Hotel (#2), share their numbers with Mahaulepu-Makauwahi Trail and Cliffs at Princeville on our list, but we could not confirm they are the same spots, so their site number is empty.

In the annual reports, Kauaʻi in March 2002 and Oʻahu in March 2003 both print 2.21 with a standard error of 0.15, the same as Oʻahu in March 2002. The 2010 report’s site values for those counts average 3.00 and 1.60, and the 2003 report’s own Table 6 averages 1.68 for Oʻahu that March. We keep both figures as printed and flag them.

The 2002 report dates its March count “26 March”, which was a Tuesday; we use the last Saturday, March 30, and say so in the note column.

The 2000 and 2001 reports describe a different count, so we leave them out: in 2000 Oʻahu observers logged every sighting from 9 am to noon, and the 2001 report describes one count day with four 15-minute scans, one on each hour. We also leave out the 2005 report, which prints its site results only as charts; the 2010 report covers the same seasons site by site.

The rescheduled 2010 count and the February 2015 sheet

The sanctuary’s March 2010 release says: “The February 27, 2010 event was rescheduled to March 6, 2010 due to the tsunami warning.” Both the March 6 and the March 27 counts appear in the press-release file, and the schedule_note column marks the March 6 rows. Neither has a row in the count-summary file, because the one release for both gives a single joint headcount, “More than 1000 volunteers”, that cannot be split between them.

We left out the February 28, 2015 results sheet’s averages. It prints 0 for every island, but its site tables show whales at most sites, so the printed averages cannot be right. No press release for that count survives in the archive copies we searched, so February 2015 has no island figure in any of our files. Its site tables are in the site-period file; our sums of them give 4.84 for Oʻahu, 5.69 for Kauaʻi and 5.48 for Hawaiʻi Island.

How to cite the data

Credit the Hawaiian Islands Humpback Whale National Marine Sanctuary as the creator of the observations, and link to the official results on oceancount.org where you can. If you used our tables, add the compilation:

Hawaiian Islands Humpback Whale National Marine Sanctuary. Sanctuary Ocean Count results sheets, press releases and reports, 2002 to 2026. Compiled by Gerrit van der Plas, sanctuaryoceancount.org (Love Hawaiʻi Guides, Inc.), dataset version 2026-09-28. https://sanctuaryoceancount.org/data/download/

If you use the 2002 to 2010 report files, also cite the reports: Maldini’s annual reports for 2002 to 2004, and Cullum, A. M. (2010), Sanctuary Ocean Count: Analysis of Humpback Whale Census and Behavior Data 2002-2010.

For registration, count dates and the official record, go to oceancount.org. To report an error in our tables, email info@lovebigisland.com with the count date and island.