A New Cornell Pavement Frost Model
By David Orr, Brendan Duffy, and Paa Mung
Since its inception in 1951, the Cornell Local Roads Program has looked at issues that are important to highway agencies across New York State examining everything from potholes on gravel roads (being more likely to form when there is no crown), to what makes the right gravel for your road; to how to meet the standards of retroreflectivity. Our latest effort examines how the road and street pavements, from Long Island to the Thousand Islands, respond to the differences in the seasons. This project, part of research originally done for the NYSDOT, is focused on helping highway and public works agencies make sure their pavement lasts as long as expected.
The inclusion of seasonal and climate issues is critical for proper pavement design, but is underutilized by many highway agencies. The most critical seasons for highways in New York State are the winter and subsequent spring thaw. Pavements are strongest in the winter, but weakest in the spring.
Knowledge of the timing of spring the Spring thaw allows highway agencies to reduce the damage caused during this season by enacting seasonal load restrictions. One primary reason for seasonality not being included in pavement design by highway departments is because it can be complicated to determine the timing of seasonal changes and determine the actual depth of frost and thaw in the pavement. There are models available to determine the depth of frost and the timing of the seasons, but many of the models can be difficult to use or understand.
An undergraduate student, Brendan Duffy, worked for Cornell for 18 months to update an older pavement frost model originally developed by the Cornell Local Roads Program that only requires a simple pavement model and daily low and high temperature data. The improved model more accurately calculated the time of the critical freezing and thawing. Figure 1 shows how the enhanced model compares to actual data for freezing and thawing in the courtyard next to our building, Riley-Robb Hall. A Transportation Research Board (TRB) paper was prepared and published in 2016 which provides the details of the model.
The pavement frost model produces accurate results at the other sites without being overly complex or difficult to use. The pavement inputs required are the pavement layers and material types, estimates of moisture content by layer, depth and value of a constant underground temperature, and knowledge of shade at the area. The only weather data required is just daily average air temperatures, which can be found on the internet from the National Climatic Data Center (NRCC) or other online sources.
A second undergraduate, Paa Mung, helped prepare statewide maps of frost depths and expected seasonal lengths for use by local agencies using data from the NRCC based at Cornell. He also calculated an analysis to show agencies the critical nature of seasonality. With the NRCC’s help, the effects of climate change are also being studied.
The effect of climate on low-volume roads
To illustrate the effect of climate, a low volume road model was analyzed with the enhanced pavement frost model. The pavement was 4 inches of asphalt concrete over 12 inches of gravel with a moderately drained silty subgrade typical for New York State. The initial traffic volume was 1,000 vehicles per day with one-eighth of the vehicles being trucks of different types. Historical weather data from 1987-2016 was used with the frost model to determine the frost depth and relative strength when frozen, thawed, or drained. The lifespan of the pavement was calculated using mechanistic-empirical pavement design methods for locations around New York State. Maps were generated using tools in ArcGIS, a commonly used geographical information system (GIS). Figure 2 shows that this thin pavement should last 10 years or more in the relatively less severe climate downstate, but would have several years less life in the harsher climate of the Adirondacks.
While a thicker pavement would not show as severe a difference, there is still a significant difference in the lifespan downstate versus upstate. A pavement that might be expected to last 25 years with normal maintenance downstate may only last 15-20 years in much of upstate New York State. There are three important things that a highway or public works agency can do to help make sure the pavement lasts as long as expected.
- Include seasonality in the original design This may include adding extra thickness in the pavement when constructed. An extra inch of asphalt concrete or 3-6 inches of gravel would be enough for the low-volume road in the model to last the 10 years expected even in the Adirondacks.
- Make sure the drainage is as good as possible This includes not only subsurface drains, but using the best quality gravel in the base. If the drainage is of excellent quality, the differences in seasonality are lessened and the pavement will last longer. Spend more on the best gravel during construction to reduce the problems in the future.
- Have a strong maintenance program Even with a modified design and good drainage, the maintenance on any highway is critical. Sealing cracks early, doing the chip seal a year earlier than typical, and being proactive with drainage maintenance can all pay strong dividends. It is a lot easier to keep a good road or street in good condition than to fix one that has broken down. Think about preparing an asset management plan.
Frost and Thaw Under the Pavement.
Figure 3 shows the expected average frost depth across New York State and surrounding areas for a relatively thin pavement. Obviously the frost depth may be shallower in a warmer winter and deeper when winter is colder and longer. Frost depth varies quite a bit across the state with the depth being only a couple of feet or less downstate to the depth being almost 4 feet across much of the North Country. Note that these depths are located under a highway where we remove the snow.
The deeper the frost, the longer the period of spring thaw as shown in the final map in this short review of pavement seasonality. Figure 4 shows the average number of days in the most critical season, spring thaw. Note that the spring thaw does not have to be contiguous. The total number of days may include January thaw as well as the expected thaw in late March or early April. Seasonal posting of highways is one possible solution, but building our roads to handle the loads may be even better.
A well maintained road is actually cheaper to drive for all vehicles due to reduced damage from hitting bumps, dips and potholes.
If seasonal posting is needed, try to target when there is a thawed layer in the pavement. Take the posting off as soon as possible. Once the initial drainage has finished a couple of weeks after thawing is done, the extra damage due to loads is mostly eliminated. Good positive drainage will make the draining period even shorter.
These three maps show the need to include seasonality in the design, maintenance and operations of roads and streets across New York State (and elsewhere). They just touch the surface of possible maps and data that can be developed.
We will put more maps on our website in the research page http://www.clrp.cornell.edu/researchprojects/seasonal.html showing additional data. Some of the maps on the website include:
- Expected number of days of winter (frozen pavement)
- Expected last day of frost in the pavement (when full draining can begin)
- Upper quartile of thawing days (a good number to use in design of a pavement versus the average)
- Average air temperature for the year
As mentioned earlier, we are also looking at the effects of climate change on pavement life. The good news is that pavement life is not likely to go down, but actually may go up. However, this is offset by increases in flooding events and a possible need to use different asphalt mixes in the surface asphalt concrete. Stay tuned.
Cornell Spring Thaw Predictor for Local Highway Agencies in New York State