Click on a question to reveal the answer. Or view all answers by clicking the button, below.
What can I use Shair for?
Shair is a modeling tool to start an environmental health dialogue among people, government, and organizations in communities.
We have two applications visualizing the model’s data – Shair Map and Shair Tech – to help users meaningfully interact with the model.
Shair Map is a browser-hosted air quality map that shows pollution exposure at scale, and down to the street-level. Individuals can use this information to assess personal pollution exposure and take action to reduce it by altering your transportation routes and changing where you spend time outdoors.
Shair Tech takes the next step by visualizing where pollution came from, and not just showing where it is. We provide source apportionment to diagnose what is causing air pollution. This data can be used to contextualize air quality, develop pollution-reduction strategies, improve emission inventories and better inform the local community.
In both apps, we provide real-time hourly pollutant concentrations and provide historical views of pollution at various time-averages (daily, weekly, monthly, seasonally, annually) to make pollution patterns visible over time.
We have scenario-running capacity and can be customized to help communities test the different pollution-reduction strategies to see what will work. This helps clarify approaches that will make the largest impact.
These are just some examples of what Shair’s features can be used for – we look forward to working with the communities we deploy in to discover new use cases.
How is your map different from other air pollution maps?
The underlying science is what differentiates Shair from other air quality solutions. We use a simplified version of CAMx, a regulatory photochemical regional model, with Shairstreet, an urban street-canyon model to estimate localized pollutant concentration. We provide results at a 10-meter resolution on a real-time basis, while visualizing pollutant source contributions.
Other air quality maps use statistical models to visualize air pollution globally.. To provide global coverage, such models maintain a coarse resolution that cannot adequately reflect how pollution affects neighborhoods on a street-by-street basis.
Alternatively, air quality sensor providers host maps of sensor measurements, visualized as discrete point measurements. These remain just points in space, without any information about the underlying processes or emissions that caused the measurement..
Instead of simply interpolating air quality measurements or using statistical modeling techniques, Shair brings together a variety of data inputs in chemistry and physics-based state-of-the-art models built in-house at Ramboll. We have crafted this solution to specifically address neighborhood-level air quality pollution on a hyperlocal level.
What does high-resolution mean? Why is it important?
Air quality models traditionally operate by assuming all data is constant within square cells of a certain size that come together to define a grid. Model output of pollutant concentration is generated for each square in that same grid. The square cell size is what defines resolution – it normally varies from 1 to 4km (or even 12km) in size. This means for a 40 km2 domain, there will be 10-40 points of input and output. Model output at this low of resolution is not particularly useful if you are interested in distinguishing between the air quality of different streets or neighborhoods, because several streets or neighborhoods are often contained within the 1-4km size of a single grid cell.
Shair’s horizontal spatial resolution ranges from the high-resolution of 200 meters to an ultra-high resolution of 10 meters. At our lowest resolution of 200m, there are still 25 times more grid points than in even the highest resolution 1km traditional model. Our 10-meter resolution produces the street-to-street view that indicates changes in air quality on the ground-level, giving our users a much more detailed picture of the local ambient air quality.
What inputs does Shair run on?
Air quality models are only as good as the inputs they use. For example, even the most sophisticated model will provide unrealistic results if the emissions provided as input are inadequate. This is the reason that significant part of the Shair modeling pipeline is dedicated to preparing and calculating input data. Shair uses many different types of data from different sources but all data fall within several broad categories outlined below. We have multiple means of acquiring the data for each category that depend on availability and customer needs.
As part of the modeling pipeline Shair runs a weather forecast model that was developed by the National Center for Atmospheric Research (NCAR), National Oceanic and Atmospheric Administration (NOAA) and several other agencies. The model is called Weather Research and Forecasting Model (WRF, the acronym is pronounced like ‘wharf’). It comes standard and configured in the Shair model.
Shair has a dedicated pipeline for calculating traffic emissions over all our modeling domains in real-time. The emissions are calculated on an hourly basis and separately for many small street sections (also called ‘links’). The emissions incorporate real-time data about congestion, which means that Shair can see rush hour traffic on local highways at different times of day as it is developing. Shair can also distinguish between different directions on the highway and show that the morning rush hour produces increased emissions in the westbound direction and the evening rush hour causes increased emissions in the eastbound lanes.
Shair uses a real-time traffic data API (application programming interface) to acquire traffic speeds, and combines them with historic counts to estimate emissions. The real-time traffic API comes standard and configured in the Shair model, but we rely on city governments or other sources to provide the historic traffic counts information.
Air Quality Monitors
Even the model with the most detailed input data and most sophisticated science will not always reflect the actual air pollution on the ground. As a check against that, we use different air quality monitors, ranging from deployed, government-managed, regulatory-compliant monitors to newer low-cost sensors. We understand the limitation of modeling as a practice, and leverage these measurements to check and correct our model. By fusing the model and measurements in such a way we provide a continuous image of the air quality over an area rather than the set of distinct data points one gets by using only monitors. While sensors do not have the same data quality standards as traditional government monitors, they can provide valuable data as long as their limitations and uncertainty are understood. Therefore, we make sure to calibrate sensor data and use both automated and manual QA/QC procedures.
Buildings and large structures affect the dispersion of pollutants. Shair uses building location and height data in a common GIS format (e.g. a shapefile or geopackage) from local governments, open data providers (e.g. Open Street Map or Microsoft’s US Building Footprints), or commercial providers. This allows us to more accurately model air quality in urban centers and similar areas with large number and concentration of built structures.
Terrain and topology can have significant impacts on the dispersion of pollutants in an area. Because of this, high-resolution terrain and land cover data are used in both our weather forecasting (WRF) and subsequent air quality modeling. This information is typically publicly available.
Traffic may be a significant source of air quality pollution in some places but it is not the only source of emissions. Shair uses emission inventories for different categories like non-road (for example, shipping, rail and air transportation), stationary, and area sources to supplement the traffic emissions that are calculated in real time. These emission inventories are usually compiled by local, state or federal government agencies.
Shair can configure and use global or regional inventories like EDGAR or the National Emission Inventory (NEI) in the US, but local inventories from municipal governments are often higher-resolution and preferable.
Why isn’t Shair available in my area now?
Shair uses scientifically-validated models with multiple complex inputs and large compute resource needs that we need to work with stakeholders across community and government to collect. Our tier 1 (‘lite’) solution is deployed across the Bay Area as of Earth Day 2020, with the full-featured version deployed in Richmond, CA via our Groundwork Richmond project, as an example of the range of our capabilities. You can browse the results of both models on our web application (https://app.ramboll-shair.com).
That said, we are working on bringing Shair to as many places as possible! Please drop us a line with your location and what air quality concerns you have so we can understand your community needs better!
How can I evaluate whether Shair could run in my area?
Shair is a model and can run anywhere in the world. As you can read above, the model is only good as the inputs – several of our inputs are standard and configured out of the box, but we need to work with local governments to gather other inputs, like building data and historical traffic data, to deliver a high-quality product. As long as there is some basic planning information available (which is often publicly available) – Shair can and will run, and will only be improved with more data points like sensors.
Does Shair need sensors? How many?
Because Shair is a model, it doesn’t need sensors to predict air quality in real-time. As long as the model has the emissions and weather data to predict concentrations, Shair can still run. For example, Shair could be a great tool to help guide where mobile monitoring laboratories should drive on a given day because the model visualizes wind and boundary layer conditions to show how expected emissions from known sources would move around and hit ground level. The same applies for siting sensors (or even more sophisticated monitors) – Shair can guide where more measurement/investigation should be done.
However, real on-the-ground observations are what take Shair to the next level, allowing the model (with emissions) and measurements to work together to estimate the best representation of the true picture of air quality at high spatiotemporal resolution. When using sensors to supplement the Shair modelling system, the higher density, the more model validation and “ground-truthing” can occur. The question of how many sensors comes down to the goal of the project or program.
I have an air quality sensor I bought. Can Shair use that data in the model?
The more data we get in our model, the better! We bring together multiple measurements, including consumer air quality sensor data via Open APIs like PurpleAir that aggregate and provide monitoring data. If you have another type of consumer sensor and would like to add the data, get in touch with us through our website – we’d love to see if there’s a way to bring it in!
What air pollution scale or air quality index does Shair use?
Shair can display maps with a variety of different scales – this can be a choice our customers make. However, for our public Shair Map, we originally made the decision to use the US EPA AQI scale (a well known scale) but for our hourly data instead of 24-hour average data which is what the AQI is meant for. Since we weren’t using the AQI in the manner it was intended, rather as a guide, we decided to use a different color scheme so as to not misrepresent our data.
A few months later, based on feedback from the community we decided to update the scale, as we describe in this blog post, to a smaller, more health-protective scale based on WHO guidelines. With the new scale the maps show more details about the concentration and dispersion of pollutants.
What does real-time mean? What does nowcast mean?
There are different definitions about what real-time modeling means in the context of air quality. Shair defines real-time as using the most recent data that constantly or periodically changes to reflect real-world conditions and constantly producing air quality results based on such data. For example, for the period of 2:00 to 2:59 PM over Richmond, CA, Shair used traffic congestion data collected just prior to the start of the hour at 1:45 PM. However, because we use weather forecasting the weather data used in the air quality model specifically matches the 2:00 to 2:59PM period. The Shair model continues this cycle operating in real-time in hourly intervals and producing a map of the air quality over an area every hour.
How are you combining static and real-time inputs?
Not all inputs to Shair are changing in real-time. Some data, that would ideally be in real-time, is actually historical data for a year or some period in the past. This is the case with the emission inventory data that Shair incorporates into the model. To use this data, Shair tries to match the most similar day in the past year to the current day. Many emission sources have diurnal, weekly or seasonal cycles, so Shair tries to find the day in the past year that most closely matches the current day and uses that data. For example, for the model run starting at 2PM on November 20th, 2019 (which was a Wednesday) we may use the data from a Wednesday in the second last week of November of a prior year. In other cases, we have surrogate data sources which change in real-time and are related to the static data. In this case, such surrogates are used to adjust the most appropriate static data in real-time and the adjusted data is used in the model for every hourly model cycle.
I’m a local industry, and your model is wrong about our impact!
This is a great opportunity to work together. We know models are only as good as their inputs, and we rely on publicly available emissions inventory data to inform our results. If there is a misalignment between your emissions and the inventory we use, we would like to work with you and other stakeholders to correct it. Workarounds could include fence-line monitoring, reviewing inventories, and other approaches. Get in touch through our website so we can work together to clear the air.
How is Shair related to Ramboll?
Shair is a Ramboll innovation company. Our founder, Julia Luongo, pitched the idea behind Shair and “graduated” from Ramboll’s internal 2018 Innovation Accelerator. The reward was funding as a separate group inside of Ramboll’s Environmental and Health Services function in the San Francisco office.
As a separate group, we are chartered with building out a new line of business in our own way, while remaining part of the Ramboll family. Our team is primarily comprised of Air Quality experts from Ramboll’s practice, and we’ve made a couple of our own hires too. We benefit from leveraging the peer-reviewed, regulatory-compliant science and expertise in Ramboll’s consulting practice in our product, while applying the different business model and go-to-market approach a product like ours requires.
How much does Shair cost?
This depends on multiple factors, including the complexity of the model, and the availability of input data. In some cases, emissions inventories have already been made available by governmental air quality agencies – this keeps costs down because our team doesn’t need to prepare the input data quite as much. Fortunately, a lot of regions have emissions inventories ready for the Shair team to implement.
Another cost-dependency is the area of coverage and resolution of data. More data generated, and stored generally leads to a higher annual subscription cost.
A lot of these questions can be more easily answered when we understand the goals you have around air quality modeling and data. Get in touch through our website, and we’ll be happy to talk through options!
Want more details?
The Shair whitepaper offers a detailed look into what we offer and our modeling methodology.
Request the whitepaper by following the link, below.