Highway sites originally modeled as part of the validation of TNM v2.5 have been re-modeled by using the specific pavement categories (DGAC, OGAC, or PCC) available in TNM v2.5 in order to evaluate the performance of TNM v2.5 when using specific pavements rather than Average pavement. Results from these models were generally comparable to the previous results which used Average pavement. The magnitude of the calibration values decreased for some sites but increased for others. These results were expected. The use of specific pavements can only improve the predicted results when Average-pavement-predicted sound levels are low for PCC sites and are high for DGAC and OGAC sites. (Site-to-site bias can produce results for Average pavement that are the opposite of this relationship, that is, high predicted sound levels for PCC and low for DGAC and OGAC.) Like many modeling decisions, it is important to use experience and measurements to guide the decision to use specific pavements. Note: At the time of this publication, federal policy requires the use of TNM Average pavement for predicting future noise impacts for federal-aid highway projects; use of pavement types other than Average in TNM should only apply to validation studies or to special programs contracted with FHWA.
Ground types were also updated for many sites based on an improved understanding of ground type classifications. These updates resulted in decreases in the difference between measured and predicted results for most open area sites (both acoustically hard and soft ground as well as near and far distances). The choice of the appropriate ground types for a site can be complex for some general ground types such as grass (refer to Section 2.2.2 for guidance). When insufficient information is available to make the most informed decision, at a future site for example, a conservative choice of ground type should be made.
Although results improved with a better representation of ground effects, for distances greater than about 500 feet, there is still some under-prediction for soft ground and over-prediction for hard ground. As a next step, predicted results from a modified version of TNM 2.5 will be evaluated to determine if further reductions in the difference between predicted and measured results at far distances can be achieved. The modified version of TNM 2.5 will incorporate a distance-dependent coherence summation between direct and reflected sound waves. Results of this investigation may result in modifications for future versions of TNM. Further research is needed in this area and will take place in the near future, contingent on available funding.
