A modified HVSR method to evaluate site effect in Northern Mississippi considering ocean wave climate

•Examines time-dependency of amplification factor with microseism source location & energy level•Presents a new data processing approach improving HVSR's estimates of amplification factor•Treats 3D spectra as constantly wandering spatial vectors forming clusters on stereographic net•Shows tighter pole clusters in the intermediate range of unconsolidated sediment thickness•Recommends the estimates of the proposed approach be based on long-term recordings

This paper presents the time history maps of horizontal-to-vertical spectral ratio (HVSR) of 12 long-term microtremor recordings (LTRs) in Northern Mississippi and the coastal area of Gulf of Mexico. These maps show that a stable predominant frequency (f0) value can be estimated across most of Northern Mississippi where the unconsolidated sediment thickness (UST) is larger than 300 m. On the other hand, estimates of amplification factor, taken as HVSR values at f0 (HVSR@f0), display a strong time-dependency, potentially caused by variations of source location and energy level of microseism, i.e. ocean wave activities in Atlantic Ocean and Gulf of Mexico. Validity of this observation is examined by calculating transfer functions between HVSR@f0 and ocean data (ocean wave height, ocean wind speed and atmospheric pressure above ocean). Additionally, 3D microtremor spectra of each LTR and those of all STRs within each 100 m-UST group are converted into spatial spectral vectors and projected on stereographic nets. Patterns of the clusters formed by these projections show that the HVSR@f0 values are related to both UST and vibration source location and energy level. Finally, a modified HVSR method based on stereographic projection is proposed to determine the average spectral vector of each cluster, which enables a more accurate estimation of amplification factor.