Youth-led research · Machine learning · Field-checked

The river is sending signals. Someone should be listening.

HydroSignal starts with research — one student's work on how streams respond to storms, including a machine-learning model trained on public rainfall and gauge data — and turns it outward: field reports in plain language, and, over time, advocacy. The point is not just to describe a storm. It is to produce understanding that informs decisions and policy, and improves people's lives — not for one season, but for the long term.

~2× / monthPlanned report cadence through spring & fall
24 hrsTarget: field visit within a day of a major storm
ML + fieldworkA machine-learning model, checked on the ground after every storm
Aerial view of the Cuyahoga River bending through Cleveland, Ohio, with bridges and the downtown skyline

THE CUYAHOGA RIVER AT CLEVELAND — ONE OF HYDROSIGNAL'S FIRST FIELD SITES

Signal snapshotSample format
LocationTinker's Creek — Bedford
TriggerStorm, ~1.8 in rain
Flow vs. normal↑ High & fast
Water clarityBrown / turbid
Field checkDebris line + bank erosion

Illustrative example of the snapshot that will open every River Report.

River Reports

One storm. One river. One clear story.

The reports are the field journal of the research — every one follows the same format (what fell, what rose, what changed on the bank, and what it means) so findings can be compared storm to storm, season to season, and eventually watershed to watershed. The three below show the launch format, from HydroSignal's first field sites in Ohio.

A muddy creek running high and brown through dense green forest after rainfall
Tinker's CreekReport No. 001 · Format example

After the Storm: Tinker's Creek Runs Brown

What a heavy overnight rain does to a small creek in six hours — flow, color, debris, and the bank that gave way near the trail.

Signal: Strong change
See how this report is built →
A waterfall in Cuyahoga Valley National Park cascading over mossy rock into a forest stream
Cuyahoga RiverFoundation piece

The River That Came Back

The Cuyahoga's recovery is famous. This background piece traces what actually changed — law, infrastructure, and decades of monitoring — and what still washes in when it rains.

Signal: Long-term trend
Explore the field sites →
A field researcher in a yellow raincoat kneeling in a stream taking a water sample with a handheld meter
Field NotesMethod

How We Take a Reading

Same spot, same photos, same checks — every time. A look at the field routine behind each report, and how public gauge data is used to verify what we see.

Signal: Method note
Read the method →

Note: These are format examples for the launch site. The first full River Report will be written by hand after a real storm, once the underlying model checks are complete.

Reading the signal

A storm, translated into three lines

This is the core HydroSignal idea: rain falls, the river answers — first in flow, then in the sediment it carries. It is also what the model learns: patterns like these, across many storms, are how a stream's behavior becomes predictable enough to plan around. Switch the view to see how each signal behaves over 48 hours.

Flow — the river speeds up

Illustrative storm pattern for a small creek. In a real report, this chart is drawn from the nearest public gauge.

Storm hits · 0h +24 hours +48 hours FLOW
Peak lags the rainFlow crests hours after the heaviest rain — the watershed needs time to collect and deliver water.
Small creeks react fastestA creek like Tinker's can rise and fall inside a day. The Cuyahoga main stem moves slower and stays high longer.
Color is dataBrown water means sediment on the move — soil, bank material, and whatever was sitting on roads and lots.
How it works

From rainfall to a report you can actually read

No jargon wall, no raw data dump. A fixed routine, repeated the same way every time, so each report can be trusted and compared.

STEP 01

A storm hits

A significant rain event over the watershed is the trigger. Rainfall totals are logged from public weather data.

STEP 02

Field visit within 24 hours

Same locations, same photo points. Flow, color, debris lines, bank condition, and wildlife signs are recorded on site.

STEP 03

The model runs — then gets checked

Ananay's machine-learning model, trained on historical rainfall and stream-gauge data, estimates how the stream should respond. Its output is tested against the field visit and the gauges. Where the model and the river disagree, the report says so — that gap is research, not embarrassment.

STEP 04

Published in plain language

One River Report: what happened, why it happened, and what it means for the people who live with the water — written for neighbors, students, and the decision-makers whose choices shape what the next storm does.

Flow

How much water is moving, and how fast, compared with normal for that spot.

In plain terms: Is the river running high, and is it still rising?

Turbidity

How cloudy the water is from sediment and runoff carried in by rain.

In plain terms: Brown water = soil on the move.

Bank & debris

Fresh erosion, undercut banks, and the high-water debris line a storm leaves behind.

In plain terms: Where did the water reach — and what did it take with it?

Temperature

Water temperature at the time of the visit, tracked across seasons.

In plain terms: Warm runoff stresses fish and stream life.

Rainfall total

How much rain fell over the watershed, and over how many hours.

In plain terms: The cause, measured before the effect.

Life signs

Fish, birds, insects, and other visible signs of stream life at the visit.

In plain terms: A healthy stream shows it.

See this research in action — read the River Reports →

The arc — a telescopic view

Research first. Then reporting. Then advocacy.

HydroSignal is deliberately built in widening rings. Each ring earns the next: no advocacy without evidence, no evidence without the fieldwork. The near view is a creek after a storm; the far view is a generation that reads its rivers fluently — wherever those rivers are.

NEAR · NOW

Research

Streamflow research, a machine-learning model, and repeated field observation at fixed sites — the evidence base everything else stands on.

NEXT

Reporting

River Reports and explainers that translate public data and field findings into language any reader can use.

FURTHER

Engagement

Talks, school and community sessions, and partnerships with watershed groups — the research, presented in person.

FAR · THE POINT

Advocacy

A public voice for streams and watersheds, grounded in documented evidence — briefs, testimony, and campaigns aimed at real outcomes: better flood readiness, smarter land use, and decisions that protect people's lives and homes not just in the next storm, but over decades.

First field sites

Where the work begins — a model built to travel

The first sites are close to home, because good fieldwork starts where you can return after every storm. The method — fixed photo points, verified public data, plain-language reporting — is written to work on any watershed, and the site list is expected to grow with the research.

Cuyahoga River · Ohio

The main stem — through Cuyahoga Valley National Park to Lake Erie at Cleveland. Slow to rise, slow to fall, and one of the great river-recovery case studies anywhere.

Tinker's Creek · Ohio

The Cuyahoga's largest tributary. Fast-reacting, heavily wooded in its gorge reaches, and the primary site for after-storm reports.

Chagrin River · Ohio

Known for cold, high-quality stretches — a useful contrast point for how different land use changes a storm's signature.

About the founder

Built by a student who studies streams

Ananay Sharda

Founder · High school junior · Streamflow researcher

HydroSignal grows out of Ananay's own research on streamflow and his interest in chemical and environmental engineering. The premise is simple: public river data already exists, but almost nobody translates it for the people who live along the water. He does — by hand, on site, after the storms that change a river most — and the longer aim is advocacy built on that evidence, for watersheds well beyond the first field sites.

  • Every report written and field-checked by Ananay
  • Model outputs verified against field observations before any claim is published
  • Drafts reviewed before publication — nothing auto-posts
  • Sized to a student's schedule: quality over volume
How HydroSignal is structured

An Ohio nonprofit corporation. Support through a fiscal sponsor.

HydroSignal was incorporated as an Ohio nonprofit corporation in November 2026. It is not yet a 501(c)(3) tax-exempt organization — its own exemption is a later step. For now, it will work with a fiscal sponsor under a grant-relationship model: the sponsor receives donations earmarked for HydroSignal and passes them through as grants, while HydroSignal keeps its own name, website, and editorial control.

That structure lets a student-led project accept tax-deductible support and apply for grants credibly, while its own tax-exempt status is pursued as a separate step.

Sponsor selection in progress — donations open once signed

What the sponsor does — and doesn't do

  • Would, once signed: receive tax-deductible donations, issue receipts, and pass funds through for HydroSignal's work
  • Would, once signed: provide the legal and financial umbrella for grants
  • Doesn't: direct the reporting — HydroSignal keeps its own identity and decisions
  • HydroSignal handles: its own reporting, site, and day-to-day operations