Sustaining tranquility in small urban green parks: A modeling approach to identify noise pollution contributors

Noise pollution poses a significant environmental challenge in Iranian urban areas, particularly within Small Urban Green Parks (SUGPs). This study investigates the extent of noise pollution in SUGPs and identifies key contributors to elevated noise levels. A total of 94 SUGPs, with a mean area of 1.86 +/- 0.99 hectares, were selected in Isfahan City, Central Iran, and subjected to 30-m Equivalent Sound Level (Lq30) measurements at their central locations. Explanatory variables were classified into two categories based on road density, traffic volume, proximity to roads, park vegetation, and physical characteristics. The relationship between noise levels and each explanatory variable was observed to be non-linear and polynomial (R-2 < 0.334) in certain instances. Our modeling results using the Multi-Layer Perceptron (MLP)-Artificial Neural Network (ANN) model (R-2 = 0.936) indicate that average noise levels in these parks surpass international guidelines, averaging 65.97 +/- 17.76 dB. Traffic volume and proximity to limited access roads were identified as primary factors influencing noise levels in SUGPs. It was observed that the design and layout of small urban parks may not effectively mitigate surrounding traffic noise, even with sufficient green barriers to absorb or block sound. Consequently, future efforts should prioritize strategic site selection for SUGPs to minimize exposure to major noise sources.