Streams are the most common fishing water in our data, and the hardest to find written up

Our first post on this dataset mentioned, in passing, that streams outnumber every other kind of mapped water and that almost none of them are individually documented. That was one paragraph in a longer piece. It deserved its own look, so I pulled all eight water-type categories in the dataset and lined up count against documentation for each one.

The short version: streams are the single largest category of mapped fishing water on earth, and they clear our notability bar less often than anything else we track, including small ponds.

The biggest category isn't a lake

Sorted by raw count, here's every water type in the WikiCatch dataset, 325,549 mapped spots across 194 countries:

Water type	Mapped spots	Share of all spots
Stream	122,217	37.5%
Lake	101,598	31.2%
River	63,073	19.4%
Reservoir	15,065	4.6%
Pond	12,699	3.9%
Sea / coastal	5,303	1.6%
Canal	3,947	1.2%
Unknown	1,647	0.5%

Streams alone are more than a third of everything mapped, more than lakes and rivers combined would be if you took away half of each. If you pictured the "typical" entry in this dataset as a named lake, that picture is wrong more often than it's right.

But almost none of them are notable

"Notable" in this dataset means a spot has cleared a documentation threshold, a linked reference or sustained in-app ratings, the same definition we've used in every post on this data. Split the same eight categories by notability rate and the ranking flips hard:

Water type	Mapped spots	Notable	Notability rate	Share of all notable spots
Stream	122,217	921	0.75%	9.3%
Pond	12,699	105	0.83%	1.1%
Lake	101,598	4,160	4.09%	42.0%
Canal	3,947	169	4.28%	1.7%
Sea / coastal	5,303	240	4.53%	2.4%
River	63,073	3,330	5.28%	33.6%
Reservoir	15,065	857	5.69%	8.7%
Unknown	1,647	120	7.29%	1.2%

Streams sit at the bottom. Not close to the bottom, at it: 0.75%, a hair under ponds and less than a fifth of the lake rate. The overall dataset average is 3.04%, so a stream is roughly four times less likely to be documented than the average mapped spot, and about five times less likely than a lake specifically.

The two numbers worth sitting with side by side are 37.5% and 9.3%. Streams are more than a third of everything mapped, but under a tenth of what's notable. Lakes run the opposite way: less than a third of all spots, but 42% of the notable ones. Whatever "gets written about" means in practice, it isn't neutral about water type. It picks lakes.

What the gap actually means in spots

If streams got documented at the dataset's overall average rate (3.04%), you'd expect roughly 3,717 of the 122,217 to be notable. The real number is 921, about a quarter of that. Put differently, there's a shortfall of somewhere around 2,800 streams that "should" be notable by the dataset's own average and aren't.

Flip it around and look at the raw count instead of the rate: 121,296 mapped streams have never cleared the notability bar. That number alone is bigger than every single fishing spot mapped in Canada, the largest single country in the whole dataset at 97,584 spots. One category of undocumented water, in isolation, outweighs an entire country's fishery.

Ponds are the other outlier worth naming here, since they finish second-lowest at 0.83%, well below the mid-single-digit rates that lakes, rivers, and reservoirs cluster around. Streams and ponds don't share a size or a use case, but they may share a structural problem: a lot of streams are short tributaries with no distinct identity of their own, and a lot of ponds are small and interchangeable with the pond one field over. Neither is the kind of water that tends to accumulate its own writeup or a following of repeat visitors leaving ratings. A named lake or a stretch of coast is more often singular enough to earn a reputation that sticks. That's a hypothesis about why the pattern exists, not something this data can prove on its own.

Methodology & honest caveats

• Source: the same water-type totals behind our first post on this dataset, 325,549 mapped fishing spots across 194 countries, each tagged by water type (stream, lake, river, reservoir, pond, sea/coastal, canal) or left as "unknown" when the type wasn't determinable.
• "Notable" = a spot with a linked reference or sustained in-app ratings (9,902 of 325,549 spots). It's a documentation threshold, not a judgment about fishing quality: an undocumented stream can easily be better fishing than a documented lake down the road.
• The "unknown" row is small and noisy. At 1,647 spots, it's the smallest category by far, and its 7.29% notability rate should be read with that in mind rather than treated as a real eighth data point next to the others.
• This measures documentation, not fishability. We don't have a way to separate "streams are genuinely less interesting to write about" from "streams are structurally harder to document" from "our documentation sources simply favor named, singular bodies of water." All three could be contributing, and this dataset can't tell them apart.
• Every number in this post is reproducible from the live dataset; nothing was hand-adjusted.

Data questions or corrections: hi@wikicatch.app. The full dataset is browsable in the WikiCatch app and on the map.