Khea Yashadhana

Warming Up “America’s Best Idea”: A Look at the Effects of Temperature on Visitation to US National Parks

Exactly a century ago this year, the US Department of the Interior entrusted the National Park Service (NPS) to protect and nurture America’s highly valued areas of national heritage. Since then, the NPS has cared for over 400 national parks, monuments, battlefields, historic sites, and more in almost every one of the US’ 3,141 counties—earning them the reputation for being “America’s best idea” because of their commitment to conserving natural and cultural resources, and contribution to the enjoyment and education of the communities they serve (National Park Service, 2016a).

As the NPS continues to grow heading into its centennial, it also recognizes that of all the global change factors the earth is subjected to, climate change and rapidly rising annual temperatures pose significant threats particularly to their designated national park units—areas that protect natural beauty and unique geological features and ecosystems. Researchers have found that climate change will not only affect species biodiversity and natural park resources, but human visitation and experience, too (Monahan & Fisichelli, 2014).

There are many factors that drive visitation to the parks, including socioeconomic constraints, geographic accessibility, population growth, and institutional schedules (like school holidays and peak vacation times), but how do regional climate patterns and temperature changes particularly have an effect on visitation?

It is valuable to seek an answer to this question considering the implications visitation timing and volume have on managing protected areas and local communities. For instance, the parks have been praised for providing employment to members of surrounding communities, and many of these jobs change seasonally depending on peak visitation (Fisichelli, Schuurman, Monahan, & Ziesler, 2015). Thus, if changing temperatures affect when peak visitation periods occur, park administrators can plan employment initiatives accordingly to meet visitor demand.

All the data visualized and analyzed in this project comes from two sources. For visitation data, the NPS Visitor Use Statistics maintains detailed records of total monthly visits to each one of its units, including 47 national parks in the contiguous United States. Data is available from 1979 to the most recent calendar year, and is obtained through counting the total number of recreation visitors who enter the parks through official park entry points. These parks in particular are a good lens by which we can analyze the effects of temperature because of its diversity in terms of climactic breadth—from deserts to forests, the tropics to the cold north.

Mean monthly temperature data was obtained from the US Historical Climatology Network (USHCN), which since 1987 has quantified national- and regional-scale temperature changes in the US with notable spatial coverage, record length, data completeness, and historical stability. Daily values for mean, maximum, and minimum temperatures are captured at each weather station in the USHCN network, and only the values that pass a quality control check are used to compute monthly average temperatures. No more than nine daily values can be missing or flagged by the QC check for the monthly average to be calculated. For our analyses, we omitted the years between 1979 and 2014 that had incomplete mean monthly temperature records. 1979 was chosen as the starting point for the analyses because it is the earliest year complete visitation data is available, and 2014 as the end because it is the latest year temperature data is available.

By putting these two data sources together and analyzing them in a variety of ways, we can assess where continued temperature patterns observed by weather stations at each park see an increase in visitation in some parks, a decrease in others, or no effects at all in some.

Of the 47 parks in the contiguous US, 31 of them have a corresponding USHCN weather station from which mean monthly temperatures are recorded between 1979 and 2014. The NPS has complete monthly visitation data for this time period, as well. A majority of these parks (23 to be exact) are located in the Western US, while the remaining nine are spread across the Midwest, Northeast, and Southeast. The map above shows the number of parks that lie within each particular state, and the shade of color shows the mean temperature of the parks in that particular state, with lighter shades being cooler and darker being warmer. Parks in Florida, for example, are very warm compared to parks in Wyoming. Now, which exactly are the warmest parks and which ones are the coolest? And what does visitation look like at those parks?

This graph breaks down the 31 parks we have data on according to their mean temperatures from 1979-2014. As we can see, Dry Tortugas is the warmest of these parks, followed by a number of parks in the Western region, as well as Everglades—the other park in Florida. The total number of visitors each park has is shown to the right of the temperature data. Visitation numbers vary greatly between the different parks, ranging from a total of just over 1M visitors at Dry Tortugas in Florida to over 329M at Great Smoky Mountains in Tennessee from 1979 to 2014.

Interestingly, the warmest park of all also happens to be the least visited, and the other warm parks also seem to have less visits relative to some of the parks towards the bottom of this graph. However, we shouldn’t be quick to conclude that extreme heat drives away visitors. Located in the middle of the Gulf of Mexico, Dry Tortugas for example is the warmest but also the most remote and difficult to access (Historic Tours of America, Inc., 2016). We should not forget that factors like geographic accessibility or population numbers around each park also contribute to visitation levels in addition to temperature. For now, let’s try to dive deeper into temperature trends over the years and see whether that’s been significant at some of the parks.

One of the most talked-about ways climate change has manifested itself is through rising annual global temperatures. Have the national parks experienced this same trend with their mean annual temperatures? Averaging each park’s temperatures for each year from 1979 to 2014 finds that six parks exhibit statistically significant (p > 0.05) changes in mean annual temperatures: Death Valley, Great Smoky Mountains, Joshua Tree, Rocky Mountain, Yellowstone, and Zion. All these parks have experienced increases, meaning on average their years are warmer as of late than they have been in the past. We will use these six parks moving forward in our analyses in hopes that they will provide us with good insights and takeaways as we explore more variables.

By selecting and comparing different parks, in this graph we can observe that the average temperatures vary from one park to the next. Joshua Tree and Death Valley, for example, are unsurprisingly warmer; Joshua Tree being where the Mojave and Colorado Deserts meet, while Death Valley is home to the lowest and driest point in North America.

For each of these parks that show a statistically significant rise in mean annual temperatures over the years, we can also plot their mean annual visitation to each year (shown in the lower half of the chart area), and find that all of them also experienced a statistically significant rise in visitors from 1979 to 2014. However, this graph does not account for population change or developments in ease of travel, which may well be confounding factors that also cause increases in visitation.

So, to further explore the relationship between temperature and visitation more exclusively, let’s take a closer look at how temperatures at these six parks fluctuate within a given year.

The visualization above has a number of elements to it. By selecting a park using the buttons on the right, we can look at the relationship between temperature and visitation for each park one at a time. The top half of the chart area plots the number of visitors in each month, while the bottom half plots the average temperature of the corresponding month. Each line represents one year, with earlier years (starting from 1979) in lighter red shades and later years (up to 2014) gradually getting darker red. Years with incomplete temperature data are omitted.

If we click through and compare the different parks, we can see that for many of them, the shape of general visitation trends map on similarly to the shape of general yearly temperature trends. For example, Rocky Mountain and Zion both see peak visits in the year happen generally during a time when the year reaches its peak temperature. However, the shape of Rocky Mountain’s visitation and temperature distribution is a narrower bell than that of Zion, suggesting that as temperatures begin dipping below 40º at Rocky Mountain, visitors probably feel that it is too cold and the park sees fewer visitors. However, temperatures at Zion remain almost steadily above 40º and what’s considered relatively ‘cooler’ temperatures at this park are still bearable for visitors, so the park doesn’t experience as drastic a dip in visitation.

Another interesting observation can be found by comparing Great Smoky Mountains with Joshua Tree. For both of these parks, peak temperatures in the year take place during the summer months—from June to around August or September—but peak visitation times differ. Visitors to Great Smoky Mountain probably find the peak summer temperatures of 70º-80º to be optimal and thus visit during this time. On the other hand, at Joshua Tree visitation actually dips in the summer, likely due to the scorching heat with average temperatures up to 90º. The 60º-80º ‘enjoyable’ range instead happens in the spring, and peak visitation is more likely to occur in the March or April months.

To get another perspective on the relationship between temperature and peak visitation periods, we can map mean monthly temperatures and visitation on a scatterplot to see when the points of peak visitation occur and at typically what temperature.

In the scatterplot above, for each park, points that lie within the grey-shaded area of the chart are points of peak visitation. By selecting and observing the parks one at a time, we can gather a number of insights. First, these peak visitation points tend to occur when the average monthly temperature is between around 65º and 80º. However, this does not automatically mean the summer months. It does for some parks like Great Smoky Mountains and Zion, where peak visitation happens in June/July, but at Joshua Tree and Death Valley they happen earlier in March/April, while at Rocky Mountain and Yellowstone they happen later towards July/August.

The differences in peak times suggest that while it may be true for some parks that visitation is greatest in the middle of the year when many folks take advantage of the convenient temperatures during periods of national holidays and summer vacation, there are also parks which experience a surge in visitation at times less popularly known as holiday/vacation periods, wherein visitors prioritize coming at points in the year when the weather is more enjoyable for those particular parks. Additionally, we may also address the idea that despite the end of year being a popular time for individuals to be off school or work and go on recreational visits, visitation numbers are consistently low during the November-January months because visitors overall prefer warmer visits across all the parks.

The trend lines in the scatterplots show that there is a statistically significant positive relationship between higher temperatures and a high number of visits at all of the parks, barring one: Joshua Tree. Could it be because Joshua Tree is extremely hot? Could there be a ‘ceiling’ temperature at which visitation trends differ before and after that point? Let’s explore this graphically.

Out of the other parks, one of the factors that sets Joshua Tree apart is its high average temperatures. As we learned in the month-by-month visitation and temperature chart above, soaring temperatures in the high-90sº become too hot for visitors to Joshua Tree. Looking at the four scatterplots above, we can better get a sense of the point in which nice and warm becomes way too hot and visitation begins tapering off not just at Joshua Tree (orange points) but Death Valley too (green points), which we know also experiences high average temperatures. For all four scatterplots, lighter shades indicate lower temperatures and darker shades indicate higher temperatures relative to each park.

At Joshua Tree, we can see that visitation actually significantly increases provided that temperatures are below 84°. Beyond this point, however, if we look at temperatures above 84° and into the 100s°, that’s when visitation begins to significantly plummet. At Death Valley, the same increase is observed for temperatures up to around 74°, before visitation then drops as temperatures increase above 74°.

I began to wonder whether other warmer parks have this ‘ceiling’ and proceeded to conduct the same analysis for those parks that had temperatures above 70°-80°. However, I found no such decrease in visitation upon reaching a certain temperature point, suggesting that this trend may be exclusive to Joshua Tree and Death Valley, or that a more sophisticated approach to analysis may be called for.

What This Means for the Parks

Though research has found that climactic conditions at many protected tourism areas are already at the warm extremes of their historical variability (Fisichelli, Schuurman, Monahan, & Ziesler, 2015), the data used for this project charts significant increases in temperatures over the years for only six national parks in the NPS system. However, these six still provide a good look into the ways in which the effects of temperature change have an impact on visitation to these parks.

We found that to some extent, visitors are most drawn to the parks at a certain range of temperature that’s commonly felt as most enjoyable, and that these periods happen at different points in the year for different parks. Beyond just avoiding freezing or heat fatigue, these optimal temperatures often mean a time when recreational experience is enhanced overall: hiking is easier, rocks aren’t covered in snow for rock climbing, evenings are less cool for camping, plants have bloomed, and spotting wildlife is certainly easier when they’re not in hibernation. Put simply, it’s a great time to go and be outside. Knowing historical temperature patterns allows visitors to plan their visits accordingly, depending on whether they seek optimal weather, less crowds, or a mix of both. It’s particularly helpful for people who, say, are located geographically far from a particularly park and need a lot of time to plan ahead for a trip.

Aside from tourists, looking at temperature and visitation can also benefit other stakeholders in many ways. Park administrators can make park operations run more effectively when they know when peak visitation seasons are going to happen and how many people on average are expected to visit. They can hire more or less staff members, see to it that visitor resources are sufficiently provided, and make sure other sources of profit like selling merchandise, food, and other paid services can meet demand while minimizing surpluses and shortages. For biologists and researchers, looking at how other biota respond to shifting temperatures and visitation timelines could be an interesting topic of research and learning, as the NPS also continues to strive for conservation amidst a rapidly changing global climate.


There are a number of limitations throughout the project that should be addressed. These include the fact that weather station data was only available for parks in the contiguous US, when one can imagine how interesting and useful it would be to observe if the temperature/visitation trends we identified hold up in diverse climactic ecosystems like Alaska and Hawaii.

Temperature data from parks that did have weather stations also presented challenges because even when all stations have data from 1979 up until 2014, there are some years in between in which mean temperature was missing for a few months—and those years had to be eliminated from the analyses completely to avoid mean calculation bias.

Additionally, the data obtained isn’t necessarily the richest because there are only really two main variables considered: temperature and visitation. As we know, there are tons of other climactic factors including rainfall, precipitation, snow, humidity, or dryness that we can safely assume also play a huge role in affecting park visitation.

In spite of this, the current project does generate a lot of potential ideas that merit future exploration into this topic of climate and visitation. For example, some research has found that peak visitation to the parks will shift earlier as spring temperatures increase (Buckley & Foushee, 2011). I tried to explore this with the monthly data that I do have, yet there was no significant shift evident, leading me to believe that perhaps visitation is shifting earlier a couple days rather than a couple months. If this were true, richer data would be needed to explore more micro level, day-to-day shifts in visitation.

Another idea would be to look at the effects of climate on other outcomes like jobs provided or financial contribution the parks bring to local economies. I wasn’t able to find monthly data on these factors, but it would be interesting to use those findings for protected areas and neighboring communities to develop adaptation strategies that can potentially capitalize on opportunities and minimize detriment related to a changing environment.


Buckley, L.B., & Foushee, M.S. (2011). Footprints of climate change in US national park visitation. International Journal of Biometeorology, 56(6), 1173-1177. doi: 10.1007/s00484-011-0508-4

Fisichelli, N.A., Schuurman, G.W., Monahan, W.B., & Ziesler, P.S. (2015). Protected area tourism in a changing climate: Will visitation at US National Parks warm up or overheat? PLoS ONE 10(6). doi:10.1371/journal.pone.0128226

Historic Tours of America, Inc. (2016). Dry Tortugas history. Retrieved from

Menne, M.J., Williams, C.N., & Vose, R.S. (2009). The United States Historical Climatology Network monthly temperature data: Version 2. Bulletin of the American Meteorological Society, 90, 993-1007.

Monahan, W.B., & Fisichelli, N.A. (2014, July 2). Climate exposure of US National Parks in a new era of change. PLoS ONE, 9(7). doi: 10.1371/journal.pone.0101302

National Park Service. (2016a). About us. Retrieved from

National Park Service. (2016b). Visitor Use Statistics. Retrieved from