Evaluation And Prediction Of Soundscape Of Urban Parks

In order to study the influence of acoustic and landscape factors on soundscape satisfaction of urban parks,three famous urban parks in Hangzhou,Lakeside Park,Taiziwan Park and Xixi Wetland·Hongyuan,were selected as research objects in this study.A soundwalk method was used to conduct field surveys of park soundscape.A head data acquisition system(Head Acoustics GmbH)and a panoramic camera(GoPro)were respectively used to synchronously acquire sound and landscape at evaluation points.A professional sound playback system(PEQ V equalizer,HDA IV.1 headphone splitter and Sennheiser HD 600 headset)and a virtual reality head-mounted display system(HTC Vive)were used to reproduce soundscape of each evaluation point in the laboratory,and laboratory evaluation of soundscape satisfaction was carried out.The software ArtemiS 10.0 was used to extract 18 acoustic indicators(sound pressure level,equivalent continuous A-weighted sound pressure level,cumulative percentage A-weighted sound pressure level,loudness level,etc.)of sound samples of each evaluation point.The on-site panoramic photos were used to extract 5 on-site landscape composition indicators(vegetation proportion,water proportion,building proportion,pavement proportion and sky proportion).The satellite remote sensing images(circular areas with a radius of 175 m centered on the evaluation points)were used to extract 30 local landscape pattern indicators,including 5 landscape elements(grass,road,building,tree,water)× 5 class-level indexes(percentage of landscape,largest patch index,landscape shape index,patch cohesion index,aggregation index)and 5 landscape-level indexes(contagion index,Shannon’s diversity index,Shannon’s evenness index,Simpson’s diversity index,Simpson’s evenness index).The correlation between mean satisfaction of park soundscape and each acoustic indicator,each landscape indicator was analyzed.On this basis,the indicators that were significantly correlated to mean satisfaction of park soundscape were selected,and a stepwise regression method was used to establish the prediction model of mean satisfaction of park soundscape.The results showed that the laboratory evaluation results of mean satisfaction of park soundscape had a consistent trend with the on-site evaluation results of mean satisfaction of park soundscape,and the former were significantly positively correlated with the latter(correlation coefficient = 0.754,p<0.01).Among all the indicators,18 acoustic indicators,1 on-site landscape composition indicator(pavement proportion),and 25 local landscape pattern indicators were significantly correlated with mean satisfaction of park soundscape.The top acoustic indicators ranked according to the absolute values of correlation coefficient were loudness level,accumulated percentage A-weighted sound pressure level(L5,L10),and equivalent continuous A-weighted sound pressure level.The top landscape indicators ranked according to the absolute values ofcorrelation coefficient were aggregation index of water,percentage of landscape of road,largest patch index of road,Simpson’s diversity index,largest patch index of tree,largest patch index of grass,Shannon’s diversity index,and Simpson’s evenness index.The stepwise regression results showed that the indicators that finally entered the prediction model of mean satisfaction of park soundscape were loudness level(LN)in the acoustic indicators,landscape shape index of road(LSI_R),aggregation index of water(AI_W)and largest patch index of water(LPI_W)in the landscape indicators.Mean satisfaction was equal to 43.180-0.110 LN-0.234 LSI_R-0.317 AI_W + 0.045 LPI_W.In the prediction model,the influence ranking(from large to small)of the four parameters on mean satisfaction of park soundscape was loudness level(standard coefficient =-0.666),aggregation index of water(standard coefficient =-0.561),largest patch index of water(standard coefficient = 0.523)and landscape shape index of road(standard coefficient =-0.310).The landscape indicators entering the prediction model were local landscape pattern indicators(class-level indexes).It could be seen that the local landscape pattern indicators had greater influence on mean satisfaction of park soundscape than the on-site landscape composition indicators.In the local landscape pattern indicators,class-level indexes had greater influence on mean satisfaction of park soundscape than landscape-level indexes.