As a HydroForecast user, you’re not only accessing cutting-edge streamflow forecasts—you’re getting a dedicated team of hydrologists, machine learning experts, and scientists committed to supporting your success. From model setup to ongoing optimization, we’re here to help you make data-driven decisions with confidence.
Here’s a look at what to expect when you start your journey with HydroForecast.
1. Discuss your basin and forecasting needs
Getting started with HydroForecast is a collaborative process. From the contract phase to model setup, we work closely with you to understand the specific conditions of your watershed and your forecasting needs. We’ll customize your model to meet your needs and fine tune it to your unique operations, including:
- Your primary use case and goals
- Forecasting horizons needs
- Forecasting site locations
- Availability of private flow gauge data
2. Share basin information and any available historical data
Your dedicated Customer Success team will guide you through the entire setup process so you have visibility into each phase of getting your forecasts operational. Collecting any available private gauge data is a critical step in the hydrologic modeling process to ensure that all relevant physical processes are captured accurately. At a high level, the pre-processing steps are:
- Delineate basins for your forecasting sites, and verify this with your team
- Incorporate management activities or other conditions impacting your basins
- Ingest any available private gauge data
Then we will run initial data quality checks. All of the data that HydroForecast uses is reviewed by our team of experts and cleaned using a set of internal tools to ensure the best performance for your specific basin.
3. Behind the scenes: Model tuning and deployment
Our team will handle the behind the scenes work to ensure your model provides accurate, reliable forecasts.
HydroForecast uses a unique, physics-informed neural network to process time-series data, delivering highly accurate forecasts that align with hydrological theory. Once we’ve integrated all relevant data, our hydrologists and machine learning experts fine tune the model to match your basin’s unique conditions. You can learn more about our theory-guided approach and the data inputs that drive our industry-leading accuracy in our blog on how HydroForecast works.
4. Access and verify the forecasts
Once we tune and validate the model, we provide users with the credentials to access your customized dashboard. After we work with you to verify the output, HydroForecast is live! Our Customer Success team will help you get familiar with the features, and you’ll be able to:
- View or download operational forecasts in your dashboard
- Integrate forecasts directly into your workflow via API
- Access weather forecasts and snow data for additional context
5. Ongoing support and optimization
Your model going live is a key milestone in our partnership, but not the end. As a HydroForecast customer, your package includes ongoing support from a dedicated Customer Success team with deep expertise in hydrology and water resources.
Have a question about your forecasts or dashboard? Reach out via our secure shared inbox and your dedicated team will help. Additional dashboard training is also available upon request.
As HydroForecast evolves, you’ll receive notifications on relevant product updates, model improvements, and new features — so your team always knows what’s changed and how to take advantage of it.
Need a deeper look at how your model is performing? Our team can work with you on performance reviews as your operations evolve. To learn more about how we evaluate forecast accuracy, check out our webinar on evaluating hydrology forecasts.
The HydroForecast team provides excellent customer service, especially with our very challenging region for modeling in Iceland. They have a truly talented and responsive team, always quick to address issues and committed to improving the forecasts.
Make sure to sign up for our newsletter to be the first to learn about new features.
