Explore the Global Wind Atlas

Use these new Living Atlas layers to assess wind energy resources for policy and planning

A Modern Wind Boom

A photo of the first wind farm in the world, built on the north flank of Crotched Mountain, New Hampshire, in 1980.
A photo of the first wind farm in the world, built on the north flank of Crotched Mountain, New Hampshire, in 1980.

The first wind farm in the world was built on the north flank of Crotched Mountain, New Hampshire, in 1980. (Photo courtesy of the Crotched Mountain Rehabilitation Center)

The power of the wind has been harnessed by humankind for millennia. First, to propel sail boats across oceans and against river currents, then by windmills for grinding grain, cutting wood, or pumping water. By the late 1800s, small wind turbines were generating electricity. It took almost 100 more years before wind turbines would be deployed on an industrial scale to capture this free and renewable resource.

Tax breaks paved the way. Following the oil crisis in the 1970s, the U.S. Federal Government offered a 15% tax credit on wind projects, which started the modern "wind rush." The first wind farm in the world was built on Crotched Mountain in New Hampshire in 1980 - but it was mostly a test of whether this network of electricity-generating turbines would disrupt the existing power grid. The good news: it was successful. The bad news: the site was lousy and it didn't really generate meaningful power. The twenty 30-kilowatt towers barely produced any electricity, so they picked everything up and headed west to windier skies.

They relocated in 1981 to a windy corridor next to I-580 northeast of Livermore, California, known as the Altamont Pass. This was the birth of the first productive wind farm in the world, which is still a fixture on the East Bay Area hillsides for passing traffic.

Photo of the wind turbines lining the green, rolling hills of the Altamont Pass, next to I-580 in the Bay Area, CA.
Photo of the wind turbines lining the green, rolling hills of the Altamont Pass, next to I-580 in the Bay Area, CA.

Wind turbines dotting the hillside along the Altamont Pass next to I-580.

In the following years, the economics of wind energy production slowly improved, as did the wind turbine technology itself. However, global wind energy production didn't become significant until the new millennium.

Wind Power Installed Capacity has seen rapid growth globally from 2000-2023. (Source:  IRENA )

With a growing need for renewable energy - and energy independence - wind farm capacity gained momentum and has more than doubled in the last 10 years. With this new interest comes a need to make more informed decisions about where to make investments in new wind farms and their supporting infrastructure, while minimizing any negative impacts to the local environment or its residents.

The Global Wind Atlas

The  Global Wind Atlas  is putting wind potential on the map - for the world to use - and is the result of a partnership between the Department of Wind Energy at the  Technical University of Denmark  (DTU Wind Energy) and the World Bank Group, with the objective of "helping policymakers, planners, and investors identify high-wind areas for wind power generation virtually anywhere in the world." Most importantly, these valuable products are now available to the GIS community in the ArcGIS Living Atlas of the World!

Photo showing a near-shore wind farm near Copenhagen, Denmark.

Near-shore wind farm near Copenhagen, Denmark.

These wind power estimates are made using a downscaling process, starting with the ERA5 climate data from 2008-2017 provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). The end products are high-resolution 250m resolution local wind estimates for the globe, at five elevations above the ground (or ocean) surface: 10m, 50m, 100m, 150m, and 200m. These different elevations account for the local "roughness" of the terrain below, and its influence on wind strength and turbulence. (More detailed methodology information can be found  here .)

Image showing the Z-dimensions available in the ArcGIS Online multidimensional tab, representing wind turbine hub heights of: 10, 50, 100, 150, and 200 meters.

Use the Multidimensional tab in the map viewer or in ArcGIS Pro to change the Z dimension (Wind Speed and Power Density) or the Display variable (Capacity Factor only).

These new multidimensional Global Wind Atlas services can now be accessed for use in analysis, planning, and visualization in ArcGIS Pro, Online, or in custom applications:

Explore the maps below for more details on these new Living Atlas services.

Wind Speed

A measure of the wind resource and is expressed as a velocity in meters per second (m/s). Higher mean wind speeds normally indicate better wind resources.

This 250-meter resolution layer displays the mean wind speed in the 10-year period (2008-2017) for five wind turbine hub heights: 10, 50, 100, 150, and 200 meters.

Click on the values below to change the Z dimension and see how the wind speed changes with height (default shown is 100 meters):

Wind Power Density

A measure of the total wind resource and takes into consideration the wind speed and elevation, or density of the wind. It is reported in watts per square meter (watts/m 2 ) and takes into account geographical variations in terrain roughness, high values of which will reduce wind power at a given height.

Compared to Wind Speed, Wind Power Density generally gives a more accurate indication of the availability of wind resources and their ability to move a turbine.

This 250-meter resolution layer displays the power density in the 10 year period (2008-2017) for five wind turbine hub heights: 10, 50, 100, 150, and 200 meters.

Click on the values below to change the Z dimension and see how the wind power density changes with height (default shown is 100 meters):

Capacity Factor

A measure of the annual energy yield of a wind turbine and is reported in megawatts (MW). Higher capacity factors indicate higher annual energy yield. The maps show estimated capacity factors, and wind turbine site suitability should be considered separately.

The capacity factor layers were calculated for three distinct wind turbines, with 100 meter hub height and rotor diameters of 112, 126, and 136 meters, which fall into three IEC Classes (IEC1, IEC2, and IEC3).

Capacity factors can be used to calculate a preliminary estimate of the energy yield of a wind turbine (in the MW range), when placed at a location. This can be done by multiplying the rated power of the wind turbine by the capacity factor for the location (and the number of hours in a year):

AEP = Prated*CF*8760 hr/year

Click on the values below to change the Display variable and see how the capacity factor changes with different IEC Classes (default shown is IEC Class 1 - 112m):

Turbine Size

The height of a wind turbine - and the size of its blades - has a lot to do with the amount of power it can produce in a given location. The height values of 10 meters to 200 meters used by the Global Wind Atlas correspond directly to common wind turbine hub heights, from a 10 meter small turbine, which might generate 200-400W, all the way up to a 200 meter offshore wind turbine with a 200 meter rotor diameter which can output 11MW, or 27,500 times more.

While many of the onshore turbines being installed are commonly in the 80m-100m hub height range, offshore turbines have migrated to 200 meters, with plans to go higher. The animation below provides some context on how large these wind turbines can get. (It's also a nice excuse to use the animated 3D turbine symbols in ArcGIS Pro!)

Wind turbines are measured according to the height of the turbine hub, not the total maximum height of the spinning blades, so measure and scale your 3D symbols appropriately.

It's best to imagine each of the Z dimensions in the Wind Speed and Wind Power Density layers as rasters hovering above the ground (or water) surface at each elevation, aligning with the center of the 5 different hub heights:

Wind Power Density values in watts/m 2  sampled on an existing wind farm in Wyoming at 10m, 50m, 100m, 150m and 200m rotor hub heights.

The Largest Wind Farms

In exploring the Global Wind Atlas, you notice some striking differences in wind resources onshore vs. offshore. Given this range of wind power, where are the largest wind projects being built right now? What are the "new frontiers" of wind energy on land and sea? Explore the map below to find where it all started, and see which countries are positioning themselves to exploit these resources in the coming decades.

Some of the largest Onshore and Offshore wind farms, by capacity, are shown on this map of Wind Power Density (100m).

The first wind farm in the world was built in 1980 on the north face of Crotched Mountain in New Hampshire.

Consisting of only 20 turbines, each producing a maximum of 30 kilowatts of electricity, it was a test bed and proving ground for future wind projects.

Lessons learned at Crotched Mountain were taken to Altamont Pass in California, just east of Livermore Valley. Wind Power Density (100m) highlights the higher energy potential in yellow and orange hues in this hilly area, now considered the global birthplace of onshore wind farms.

Onshore wind energy has been dominated by China, which had 404,605 MW of installed electricity capacity in 2023, according to the International Renewable Energy Agency (IRENA). By comparison, the United States had less than a third of that, at 147,979 MW.

Not only the largest, but the top 3 onshore wind farms in the world can all be found here - and China is continuing to greatly expand their capacity in the coming decade.

Chart showing the increase in China's onshore and offshore on-grid electricity installed capacity from 2000-2023.

Onshore and offshore on-grid installed wind farm capacity in China. (Source:  IRENA )

The Gansu Wind Farm is located in the Gobi desert, and includes solar farms, merging their power grid infrastructure.

Getting the electricity to the port cities that could use it - nearly 1,600 km away - was a challenge early in the project. It wasn't until 2017, 8 years after the first wind turbines were built, that the 2,383 km long Jiuquan-Hunan HVDC transmission line entered service and Gansu's production capacity could be fully utilized.

When it comes to offshore wind farms, 5 of the top 6 by installed capacity can be found off the coast of the United Kingdom, in the North Sea. (100m Wind Power Density layer shown.)

UK offshore electricity capacity is on track to overtake onshore in the coming years. (Source:  IRENA )

A simple comparison of the Wind Power Density at each of these 12 locations and 5 turbine heights is easy to do using the multidimensional analysis tools in ArcGIS Pro, including  Sample  or  Zonal Statistics . Interact with the chart below to sort power density by turbine height at each of the 12 largest wind farms.

Offshore farms, free of surface obstacles to wind currents, can produce significantly more electricity than those onshore, with larger turbines.

Sampling locations on a map to determine wind power potential is a powerful capability, but many other criteria are weighed when new wind farms are planned and constructed. In the final part of this story, let's use these Global Wind Atlas layers to determine some optimal candidate locations for a new wind farm.

Wind Suitability Model

These new Living Atlas wind layers are more than just pretty maps - their primary purpose is to help make better-informed decisions on where to invest in future wind farm infrastructure, be it onshore and offshore. To explore this analytical process, we'll use a particularly wind-rich state for a deeper dive: Wyoming. Here are some Wyoming wind stats:

Imagery map showing the outline of Wyoming, the top wind potential state in the contiguous United States.

The 9th largest state in the Union: the Equality State, the Cowboy State, Big Wyoming.

  • Of the lower 48 states, Wyoming is the windiest, with an average wind speed of 21.5 mph (Alaska is first of the 50, at 21.9 mph).
  • Wyoming produced about 3,100 megawatts of electricity in 2023, about 15 times the amount of electricity it consumes.
  • National Renewable Energy Laboratory (NREL) data shows that over 50 percent of the best quality (Class 6 and 7) wind capacity in the continental United States is located in Wyoming.

Wyoming's bounty of wind potential is a byproduct of the state's geography: as prevailing westerly winds from the Pacific pass over the Rocky Mountains, they funnel through mountain passes and gust over the plains to the east. With these attributes, Wyoming makes for a great case study to determine some ideal candidate sites for future wind farms.

Wyoming has unique geography, vast wind potential, and natural resources worth protecting.

Suitability Modeling

Suitability Modeling always starts with a goal in mind. In our case, it is to discover suitable locations for one - or many - wind farms. It is an iterative process where you determine and compile the relevant base criteria to make the most informed decision, then analyze, transform, weight and combine these layers to arrive at the most suitable locations. By design, it is an iterative process, where subject matter expertise meets experimentation, evaluation, and repetition.

Infographic showing the factors which contribute to selecting wind turbine sites, within the framework of the Suitability Modeling workflow in ArcGIS Pro.

Suitability criteria used for offshore wind farms differ from those used onshore.

Finding potential wind farms is a perfect use case for the  Suitability Modeler  tools in ArcGIS Pro. For this exercise, we'll use the Wind Power Density layer to determine where in Wyoming we should place a new wind farm, using several Living Atlas layers to aid in our site evaluation and selection.

This exercise will be specific to locating onshore wind farms; if the site selection was for an offshore farm, the criteria would look quite different, as illustrated in the diagram to the right. While both locations share a need for strong wind, our terrain, transportation, and other infrastructure layers will be of the terrestrial flavor.

Suitability Criteria

The first step in the modeling process is to compile and prepare the layers that will participate in the modeling process - which are most relevant to decision-making. In this case, some are related to the availability of wind resources, which include Power Density and the types of landforms where wind is highest. Other criteria might be more concerned with how difficult it will be to access and construct the wind farm, or connect it up to the power grid. This can be done using  suitability submodels , in ArcGIS Pro, but for the sake of simplicity we'll stick to a single model. (For a deeper dive, read the  Multi-objective modeling with submodels using the Suitability Modeler  ArcGIS blog).

Once you have a list of criteria assembled, the layers then need to be prepared, which may include creating derivatives of the source data or projecting to a common coordinate system amongst all criteria (geographic coordinate systems using decimal degrees are not allowed). For our case study, we'll use these 5 wind farm criteria, in Web Mercator Auxiliary Sphere for simplicity. Preparation notes are provided for each:

  •  Global Wind Atlas - Power Density 
    • Used the 100m turbine height, approximating the ~80-94m wind turbine sizes that are common in Wyoming, as seen in the Living Atlas US Wind Turbine database.
  •  Elevation : Slope
    • A slope raster (in percent) was created from a statewide Wyoming digital elevation model (DEM). Construction on steep slopes is dangerous and expensive, so we are going to prefer low values (up to ~10-15%) and not build if it's much steeper than that.
  •  Elevation : Landforms
    • The  Geomorphon Landforms  tool was used to convert a statewide Wyoming DEM to a classified landforms raster, delineating slopes, valleys, peaks, ridges, and other landform types. You will often see wind farms on ridges and flatlands, but never in hollows or pits.
  •  USGS Wyoming Roads 
    • Road segments with road class of 1, 2 and 3 were used to represent just major roads in the state, which can support the enormous blades and towers on their journey to assembly on the wind farm site.
    • The  Distance Accumulation  tool was used to create a raster surface whose cell values represent the distance to the nearest road.
  •  U.S. Electric Power Transmission Lines 
    • The  Distance Accumulation  tool was used to create a raster surface whose cell values represent the distance to the nearest power transmission line.

Wind-related suitability layers derived and projected for use in the model.

Knowing where you can't place a wind farm can be just as important to know as where you can, which is where restricted areas come in.

Restricted Areas

Whether it's a construction restriction due to land ownership, line-of-sight or viewshed analysis, or areas where plant or animal habitats are endangered or imperiled, these locations are taken into account early in the suitability modeling process - effectively masking and removing them from contention. You will find all of these in the Living Atlas:

Restricted areas where wind farms should not be built, including on existing wind farm sites.

Configuring the Suitability Model

A real power company constructing a real wind farm would likely have a much more exhaustive list of suitability criteria and restrictions, but this case study will illustrate the overall process with a simplified set of inputs and configuration steps. Given the same starting datasets, 10 people would come up with 10 very different suitability models, so do not take any of this author's choices along the way as "the truth," but rather, a series of semi-informed estimates with the end-goal of "that seems reasonable."

With that important disclaimer out of the way, use the slide show below to take a tour of the process:

Start a new Suitability Model

Add all the prepared criteria and restriction layers to a new map in ArcGIS Pro.

In the Analysis tab of the top ribbon, select Suitability Modeler.

Screenshot of ArcGIS Pro showing Contents pane with restriction and suitability layers., with map centered on Wyoming.

Configure Settings

In the Suitability Modeler pane, specify a new model name and configure the input type, scale, and weight options. (We're using default values.)

Next, add the Restricted Locations. We've pre-processed and merged the national parks and biodiversity areas into a single restriction raster, but you can also use the Add Clause button to create an "OR" expression to combine multiple restriction layers.

Click Apply and a new Restricted areas group layer will be added to the Contents pane on the left.

Screenshot of ArcGIS Pro showing input of restriction layers into the Suitability Modeler with map of Wyoming.

Add Suitability Criteria

Click on the Suitability tab in the Suitability Modeler pane and use the drop-down menu to select multiple inputs, in our case, these criteria:

  1. Slope
  2. Landforms
  3. Road Distance
  4. Transmission Lines Distance
  5. Global Wind Atlas - Power Density (100m)

Once selected, click the Add button to add them to the model.

Screenshot of ArcGIS Pro showing Suitability Modeler criteria input dialog with map of Wyoming.

The new criteria layers will be added to the Suitability Model group in the Contents pane on the left. Now we can start to configure each input.

Click the radio button next to the Global Wind Atlas layer in the Suitability Modeler pane to open the Transformation Pane.

This process will transform the min-max range of the data in the raster to output values between 1 and 10, representing least- to most-suitable regions.

Configure Wind Power Transformation

With the Global Wind Atlas layer selected, the Transformation Pane will open below the map.

Because our data is continuous, we will use a function to transform the value range to a suitability range. Use the drop-down to select a function that is appropriate to the preferred distribution of the values in your raster, using the graph on the right.

List of Continuous Functions available in the Suitability Modeler in Pro.

We've selected the MSLarge function, which prefers the highest values. Feel free to experiment with the Mean multiplier and other options to find a curve shape that best ranks the suitability of the wind power. As you change the settings, the Suitability_map layer in the Contents will update, giving you immediate feedback on your choices. (green = most suitable, red = least suitable)

It can be helpful here to compare the Power Density values to the Suitability map to make sure high wind areas correspond to the green high suitability areas.

GIF animation alternating between wind power density and its transformed suitability layer.

When transformed, green areas with high suitability match the brighter locations with high wind power density.

Configure Transmission Lines Transformation

Next, select the Transmission Lines Distance layer in the modeler window to configure its transformation.

Here, we are using a Continuous Function again, preferring small values, with a Mid point of 20km and a maximum Upper threshold of 50km. This will limit the distance we are willing to build new power lines to connect to the existing electrical network.

As you move through the modeling process, it can also be helpful as a check to add the existing wind turbine points to the map to make sure we are capturing the existing wind farms (which were presumably constructed under similar constraints) in the green areas of the Suitability_map.

Configure Road Distance Transformation

We will treat this function the same as transmission lines, as they're both infrastructure and have similar building/distance constraints to keep costs down.

Only the larger, multi-lane classes of road segments were used in this analysis, due to the challenge of transporting large wind farm equipment on small or single-lane roads.

Photo of a large turbine blade being trucked through a small town.

Transporting large wind turbine blades requires advanced planning to traverse the road network.

Configure Landforms Transformation

Next, configure the Landforms layer - this time, we are dealing with a categorical raster with 10 landform types, so select the Unique Categories tab in the Transformation Pane.

The drop-down allows us to change from numeric values to text Landform names. Once populated, the Suitability values in the right column can be ranked from most suitable (10) to least suitable (1).

Initially unsure how to rank landforms, one option is to find out on what landform type each of the existing 1,560 wind turbines in Wyoming were built. Using the existing distribution of wind turbines to set landscape type ranks falls back on the expertise of those who have already paid to construct these farms in Wyoming. We can mimic their distribution in our suitability model.

Flat, Shoulder, and Ridge are the most popular landform types for installing wind turbines in Wyoming.

Flats were given the top ranking of 10, followed by Shoulder at 8, down to the least desirable Hollow and Valley at 1.

Configure Slope Transformation

Click on the Slope layer in the modeler to configure the final layer transformation.

We are back to a continuous function, MSSmall. This will prefer small values (a lower slope % = flatter) as it is expensive and dangerous to construct wind turbines on steep slopes. Here we can use the Mean multiplier and Threshold values to make the function curve drop down dramatically to 1 before we exceed 30% slope, with the bulk of our preference in the tall 0-4.4% green bar in the graph to the right.

As always, it is useful to turn off the Suitability_map in the Contents pane to check the Transformed Slope surface against the exiting wind turbines. If they're mostly in the green areas, we're probably making reasonable decisions based on previous construction restraints.

Add Criteria Weights

Once all the criteria are transformed, each one is then weighed against the others. These weights then multiply each individual input's contribution to the final suitability model.

If one criteria has a weight of 1 and another has a weight of 2.5, the second is 2.5 times more important.

In our case, the most important criteria is Wind Power Density, so it received the highest weight value: if there is no wind, there is no wind farm. Landforms were given the lowest weight, with the other criteria falling in between. (These weight values are for demonstration purposes only.)

With weights assigned to all criteria, click the Run button to calculate the suitability raster.

Locate Suitable Regions

The final step is to combine the suitability layers, transformations, and weights to determine where we can place a new wind farm. Click on the Locate tab of the modeler pane to enter the tool parameters.

For this case study, we'll look for 10 suitable wind farm sites, each about twice as big as the largest cluster of wind turbines in the US Wind Turbine Database - about 500 square kilometers. That gives us a Total area of 5000 square kilometers, which goes in the first box.

We'll use the default Circle region shape, and specify a Shape/Utility tradeoff of 25%, which will prefer higher suitability over maintaining our target shape.

This is also where we can add an existing regions feature to our model, which are our existing wind farm points that we don't want to build on. The  Find Point Clusters  tool makes this easy, using the Self-Adjusting (HDBSCAN) option, with 10 features per cluster, followed by calculating the  Minimum Bounding Geometry , using the Cluster ID to create individual wind farm polygons.

GIF animation showing how Find Point Clusters tool can be used to create cluster polygons for existing wind farm areas, for input as exclusion areas in the Suitability Modeler tool.

Existing Regions: individual wind farm clusters are detected, then bounding polygons are created according to the Cluster IDs.

Once the inputs are specified, click on the Run button to create the suitability raster dataset.

Explore the Results

When the calculation is complete, the 10 suitability regions are added to the map. Not too bad for a first run - the suitable regions are nearby existing wind turbine locations, which is encouraging.

But maybe the regions are a little too close to other farms - to reduce that encroachment, we can go back and re-run the model, this time providing a Minimum distance between regions value. Or buffer our wind turbine cluster polygons. We might also want to try a slightly larger Region maximum area to place a larger wind farm. Or add new criteria. It's an iterative process!

To summarize the work we've done, use the Generate Report button for a PDF summary of the model, or to access a suite of analytical maps and tools, use the Evaluate tab in the Suitability Modeler pane. If you are interested in a deeper dive, go  here  for more resources.

With a suitability model completed, use the map below to investigate the 10 optimal wind farm sites that were generated. Existing wind farm points have additional attributes about the year they were installed and the height of the wind turbine hub - useful for creating compelling 3D visuals.

Suitability regions, existing wind turbine farms, and restricted (biodiversity and national parks) areas in Wyoming.

The Living Atlas was created to put authoritative geographic data in the hands of people who can make a difference through its application towards a better and more sustainable world. Coupled with powerful visualization and analytic tools in ArcGIS, the Global Wind Atlas is a valuable resource to help achieve this goal.

Resources

We're excited to add these new Global Wind Atlas resources to the Living Atlas for the renewable energy and GIS community. Additional questions or feedback? Feel free to contact the  Living Atlas Environment Team! 

Finally, please use the links below to access the datasets and blogs referenced in this story.

The first wind farm in the world was built on the north flank of Crotched Mountain, New Hampshire, in 1980. (Photo courtesy of the Crotched Mountain Rehabilitation Center)

Wind turbines dotting the hillside along the Altamont Pass next to I-580.

Near-shore wind farm near Copenhagen, Denmark.

Use the Multidimensional tab in the map viewer or in ArcGIS Pro to change the Z dimension (Wind Speed and Power Density) or the Display variable (Capacity Factor only).

The 9th largest state in the Union: the Equality State, the Cowboy State, Big Wyoming.

Suitability criteria used for offshore wind farms differ from those used onshore.

Onshore and offshore on-grid installed wind farm capacity in China. (Source:  IRENA )

When transformed, green areas with high suitability match the brighter locations with high wind power density.

Transporting large wind turbine blades requires advanced planning to traverse the road network.

Existing Regions: individual wind farm clusters are detected, then bounding polygons are created according to the Cluster IDs.