a road closed barrier weighed down by a sand bag on a flooded roadway

Adobe Stock | Ajax9

Monthly High Tide Flooding (HTF) Outlook: Experimental Product

High tide flooding is becoming more frequent along U.S. coasts, and because it depends largely on predictable tides, it can be anticipated weeks to months ahead. This page shows experimental next-generation forecasts of high tide flooding at 122 NOAA tide gauges for leads of one to six months, developed by NOAA's Physical Sciences Laboratory and CIRES in partnership with NOAA CO-OPS.

The map below summarizes the lead-month-1 forecast: each gauge is colored by the likelihood of at least one flood day during the month. Daily probabilities and expected flood-day counts for each station are available under the High Tide Flood tab.

latest map

About these forecasts High tide flooding is minor coastal flooding that happens at high tide, even without a storm. It can close roads, damage infrastructure, and disrupt coastal communities. NOAA’s National Ocean Service, through its Center for Operational Oceanographic Products and Services (CO-OPS), issues the official Monthly High Tide Flooding Outlook. It gives daily flood chances up to a year ahead. Communities use it to plan road closures, protect electrical grids and other infrastructure from saltwater damage, and reduce costs to local economies. The forecasts on this page are an experimental, next-generation version of that outlook, looking one to six months ahead. NOAA’s Physical Sciences Laboratory (PSL) and the Cooperative Institute for Research in Environmental Sciences (CIRES) developed it with CO-OPS, and it is being evaluated for possible use in the official product. The goal is more accurate flood probabilities, especially for the highest water levels and at stations where natural climate swings, such as El Niño, strongly affect sea level.

How the forecasts are made We treat the water level at a tide gauge as the sum of five pieces:
  1. The predicted tide, the regular rise and fall caused by the moon and sun
  2. The long-term trend, such as gradual sea level rise
  3. The seasonal cycle, the usual highs and lows through the year
  4. The monthly sea level anomaly, how much higher or lower than usual a month’s average water level is
  5. Short-term ups and downs, lasting from hours to weeks
The monthly anomaly is forecast by combining several models (a “multi-model ensemble”) and calibrating the result. It currently uses Environment and Climate Change Canada’s GEM5.2-NEMO seasonal forecast model and a data-driven model called CS-LIM. The short-term ups and downs tend to push water higher more often than lower, and they tend to last from one day to the next. A statistical model fitted to each gauge’s hourly record represents them. Putting the pieces together produces a large set of possible hourly water levels. The chance of flooding on a given day is the share of those possibilities in which the water goes above the flood threshold at any hour that day.

What is new Compared with NOAA’s current high tide flooding outlook, this system:
  1. Predicts the monthly anomaly with a calibrated mix of physics-based and data-driven seasonal forecasts, instead of assuming today’s anomaly gradually fades (“damped persistence”).
  2. Uses a lopsided (skewed) distribution for hour-to-hour variability, rather than a symmetric bell curve (Gaussian), so large rises are treated as more likely than equally large drops.
  3. Calculates daily flood risk directly from the simulated hourly water levels.

How well do the forecasts work? In tests on past years, the new forecasts were more skillful than the current outlook at most stations, at every lead from one to six months. They also beat a baseline of typical conditions for the time of year (a “seasonally varying climatology”) at about 70 of 116 gauges, compared with about 63 for the current outlook. The tests re-ran the forecasts for 1993–2020 and scored them only on years not used to build the model (10-fold cross-validation), with the long-term trend removed, for a high flood threshold set for each station. The forecasts do best on the West Coast, in Alaska, and in the Pacific Islands, where monthly sea level anomalies are easiest to predict. On the Atlantic and Gulf coasts, most of the skill comes from the tide. Tests in which the monthly anomaly was known perfectly show that substantial improvement is still possible, especially on those coasts.


Disclaimer These forecasts are experimental.
NOAA/PSL and CIRES/University of Colorado are not responsible for any loss occasioned by the use of these forecasts.

High Tide Flooding: Station Plots

Experimental Product showing "High Tide Flooding" predictions at ocean stations in the US.


dynamic plot

Monthly High Tide Flooding Outlook FAQs

Experimental Product

What does this outlook tell me?

The outlook shows the daily likelihood that water levels will exceed a selected flood threshold at a NOAA tide gauge, one to six months in advance. It provides daily flooding probabilities and estimates of how many flood days may occur during the selected month.

How can my community use this information?

The outlook can help you plan staffing, coordinate resources, and prepare outreach for periods with greater flooding potential. As those dates approach, National Weather Service forecasts and alerts give near-term guidance for response decisions. During an event, NOAA’s Coastal Inundation Dashboard can be used to monitor current water levels.

What does “forecast lead” mean?

Forecast lead describes how far ahead a forecast looks from its initialization date, the day the forecast starts. For example, a forecast initialized on November 1 has a forecast lead of one month for November, two months for December, and so on. Forecasts with longer lead times made at different times may give slightly different estimates for the same month.

Check the heading above the station plot to confirm both the initialization date (the day the forecast was started) and the lead time (the month being forecast).

What is a flood threshold?

A flood threshold is the coastal water level height where flood impacts (e.g. water reaching places it normally doesn't) begin to be observed. The outlook offers the following flood thresholds:
  • National Weather Service (NWS) thresholds are established by local weather forecast offices through working with local emergency managers to determine the elevation when a given area begins to be affected by flooding.
  • National Ocean Service (NOS) thresholds provide nationally consistent flood levels above the average local tide range, derived from NWS thresholds. They can be used to compare multiple locations or can be used where NWS flood thresholds are unavailable.
  • 0.5 - 2.5 ft above MHHW thresholds provide a customizable option for users in areas that may flood before water levels reach the height of the established NWS and NOS thresholds.

For NWS and NOS thresholds, minor flooding is disruptive, shallow flooding with a low threat of property damage and no direct threat to life.

All flood thresholds on the outlook are shown relative to Mean Higher High Water (MHHW), which is the average height of the higher high tide of each day.

What is the difference between “Minimum Flood Days Per Month” and “Expected Number of Flood Days”?

These measures answer two related questions: “What is the chance of at least this many flood days?” and “How many flood days are expected overall?”

“Minimum Flood Days Per Month” shows the chance of having at least a given number of flood days during the month. For example, a 61% chance of at least four flood days means that four or more days would flood in roughly six out of ten similar situations. The chance of at least one day is always higher than the chance of at least four days.

“Expected number of flood days” summarizes the forecast as a single estimated count, taking all possible outcomes and their probabilities into account. For example, 3.9 means about four flood days are expected on average. The actual count could be higher or lower, and four is not necessarily the most likely individual outcome. The number can include a fraction because it is an average.

Neither measure identifies which days will flood. To see which days have greater flooding potential, refer to the daily probabilities in the calendar.

Where do I find monthly and daily flood-day probabilities?

The map shows the chance of at least one flood day during the selected month (colored dot) and the expected number of flood days (number next to the dot). Once you choose a station, the calendar shows the chance of exceeding the selected flood threshold on each individual day, as well as the chance of at least one flood day, at least two flood days, and so on through at least five flood days.

How does the outlook account for El Niño and other ocean conditions?

The outlook combines predicted astronomical tides, long-term sea level trends, seasonal changes in average water levels, forecasts of monthly sea-level anomalies, and shorter-term water-level variability. A sea-level anomaly is the difference between observed water levels and predicted astronomical tides; it can be positive or negative.

The outlook can account for El Niño’s influence through seasonal prediction models that forecast changes in monthly sea levels. For example, El Niño can elevate sea levels along the U.S. West Coast, making it more likely that high tides will exceed flood thresholds.

How do I know if the predictions are accurate?

You can assess how well the predictions perform by comparing forecasts with observations, understanding the forecast uncertainty, and considering the available evaluation record. Performance is best judged over many days and months rather than individual events.

Compare forecasts with observations. Forecasts are checked against observations in near real time. You can explore these comparisons in two places, using preliminary observations:

  • Calendar: Outlined days indicate that water levels exceeded the selected threshold.
  • Plot on left: Black dots connected by a dashed line show observed daily maximum water levels. Dots above the selected flood threshold (the red dashed line) indicate flood days.

Understand forecast uncertainty. The gray band in the left panel shows the 99% confidence interval, or the range that 99% of model simulations fall within. If the forecast is working well, observed daily maximum water levels should stay inside the band about 99% of the time.

Daily flood probabilities also express uncertainty. Because the daily flood probabilities are chances rather than yes-or-no calls, a flood on a low-probability day, or no flood on a high-probability one, does not by itself mean the forecast was wrong.

Consider the available evaluation record. A longer record allows for a more thorough assessment of performance across different conditions. The model was built using data through 2020, and real-time forecasts have been saved since August 2024, allowing you to check how well past outlooks performed. More detailed performance evaluations are planned for the webpage.


Forecast updates: Forecasts start from conditions on the 1st of the month and are posted around the 15th, once all input data become available. Questions or feedback: Contact matt.newman@noaa.gov with questions, feedback, or technical issues.
Document Last Updated: November 2026

Technical Details Coming Soon!

Skill Assessment page to come

NMME SSH Skill

        
  Linear Trend Removed?  no   yes    Anomalies   Probabilities

Forecasts/Verifications

This page is under development and will be released soon.

Rise Time Series





Anomalies   Probabilities

RISE Project

Results of the RISE project are described in this paper:

  • Long. X., M. Newman, S.-I. Shin, M. Balmeseda, J. Callahan, G. Dusek, L. Jia. B. Kirtman, J. Krasting, C.C. Lee, T. Lee, W. Sweet, O. Wang, Y. Wang, and M.J. Widlansky, 2025: Evaluating Current Statistical and Dynamical Forecasting Techniques for Seasonal Coastal Sea Level Prediction. J. Climate, 38, 1477–1503, doi: 10.1175/JCLI-D-24-0214.1.

Details of the techniques

NMME:

For full, more technical descriptions, see:

Linear Inverse Model (LIM):

For full, more technical descriptions, see:

JPL ECCO adjoint:

This hybrid dynamical method is based on the mathematical convolution of atmospheric forcing with the lead-time dependent sensitivities of sea level to atmospheric forcing. The sensitivities of sea level to atmospheric forcing are computed by the ECCO adjoint model with the observationally constrained ECCO ocean state estimate as the background ocean state. The atmospheric forcing is based on the combination of ECCO forcing (constrained by observations) prior to prediction time and ensemble NMME predictions of atmospheric forcing after prediction initialization time with bias corrections applied using ECCO forcing climatology that is observationally constrained. The convolution integrate in space, lead time, and forcing type give predictions of sea level anomalies.

Decomposition of the convolution in space, lead time, and forcing type allows us to examine the sources of uncertainties for the sea level prediction.

For full, more technical descriptions, see:

Statistical downscaling:

Currently, the large computer models that form the basis of seasonal climate prediction systems produce coastal sea level forecasts spaced about 100 km apart. This is too coarse to meet the needs of U.S. coastal ocean management and services, which are becoming increasingly important as sea levels rise in a warming climate. Here we use a method to generate monthly sea level predictions on distances as small as 10 km apart, by applying the observed statistical relationship between sea level variations on scales of 100–1,000 km and finer-scale coastal ocean observations to the original coarser model predictions. For the RISE project, this was applied to the SEAS5, CCSM4, and SPEAR hindcasts.

For full, more technical descriptions, see:

Downloads

Individual map files are also available through OPeNDAP. File names are structured as

https://psl.noaa.gov/thredds/dodsC/Projects/RISE/webData4OPeNDAP/DATANAME/DATANAME_ssh_YYYYMM_LagL.nc

where
YYYYMM = initialization date
DATANAME is the dataset/model forecast name id (CCSM4, CFSv2, CanCM3, CanCM4, GFDL_FLORB01, LIM, NMME_MEAN, AVISO, ORAs4)
L is the forecast lead in months (ranging from 0 to 12)

More information on PSL OPeNDAP

Note that verification data are stored similarly to model hindcast data. For example, the ORAs4_ssh_201001_Lag6.nc file contains the observed ORAs4 SSH anomalies used to verify the Month 6 forecast initialized January 2010 (i.e., the July 2010 observed anomaly).

All maps are on a 1 deg x 1 deg grid, interpolated/regridded as necessary.

This is a Research and Development Application