Florida Noise Barrier Evaluation and Computer Model Validation

Abstract

The results of a project that investigated the effectiveness of in situ noise barriers in Florida are presented. The prediction accuracy of the FHWA Traffic Noise Model (TNM) is compared with STAMINA 2.0 and 2.1 (Florida-specific). A total of 20 barrier sites were visited during a 3-year period that resulted in 844 discrete 20-min equivalent sound level (L_eq) measurements behind the barriers. Barrier insertion loss was determined using the ANSI indirect barrier method. A methodology was developed to estimate shadow zone length created behind highway noise barriers. All of the barriers tested were effective (>5 dB:L_Aeq insertion loss at distances equivalent to the first row of homes, where L_Aeq is the A-weighted L_eq) except one site because of marginal additional shielding from a berm–barrier combination. Only three sites had an insertion loss of less than 5 dB at distances representative of the second row of homes. Overall, measurements indicate that the barriers provide substantial sound level reduction for residents along the highway. TNM was the best prediction model when considering all test sites; however, the STAMINA models were more accurate at predicting source level. TNM predictions using the Average pavement input overpredicted the reference sound levels measured at these sites. TNM predictions using the OGAC (open-graded asphalt concrete) input were improved (under 2 dB:L_Aeq of error) over those using the Average pavement type input. This result is expected because Florida uses an open-graded asphalt friction mix.