Abstract: 【Objective】The objective of the present study was to compare the different ultrasonic acoustic emissions(UAEs)indices and signals extraction in relationship with the phenomena occurring during the dehydration of tree stems from in situ-detected UAEs to obtain a better quantification of embolism by application of non-destructive UAE techniques.【Method】The characteristics of the cumulative AEs indices(i.e., count, energy and amplitude), and the relative UAE rate distributions were analyzed at four amplitude peak ranges(35-40, ≥40-50, ≥50-60 and ≥60 dB), and all the UAEs were monitored with the Micro-II Digital AE System for Larix gmelinii and Pinus koraiensis in Northeast China. The end points of the three indices for five groups of signals were determined by the end of the maximum relative UAES rates, which mathematically corresponded to the local maximum of the third derivative of the curve of cumulative UAEs versus time, and filtered out the interference signals at late stages of dehydration. The P₅₀ values, which are indicators of xylem embolism vulnerability, were calculated based on five groups of unfiltered and filtered data.【Result】① The cumulative curves for the indexes at ≥40-50 and ≥50-60 dB were shifted left 0-5 h in L. gmelinii and 10-25 h in P. koraiensis, whereas the curves at 35-40 and ≥ 60 dB were shifted right 5-35 h in L. gmelinii and 45-60 h in P. koraiensis as compared with those of the total signals. ② The relative UAE rate distributions at ≥40-50 and ≥50-60 dB for the two species were basically consistent with the corresponding total signals. Moreover, the maximum relative rate of energy was 2/5-3/5 of the values for the total signals, which was markedly higher than that at 35-40 dB and ≥ 60 dB. ③ In L. gmelinii, no significant differences were found in the P₅₀ among count, energy, and amplitude, based on unfiltered or filtered total signals, which ranged from -2.92 to -2.96 MPa, and from -2.73 to -2.74 MPa, respectively. For P. koraiensis, P₅₀ calculated from the total signals of three unfiltered indices showed no significant difference, and ranged from -3.60 to -3.84 MPa. However, values based on filtered total signals of energy were significantly higher than those of the filtered amplitude, which ranged from -3.40 to -3.83 MPa. ④ P₅₀ were higher in the three filtered indices than in the unfiltered data for each of the two species. The values at ≥ 60 dB based on unfiltered or three filtered indices had the largest difference in the total signals. Moreover, the values for the rest of the amplitude peak ranges showed no significant difference in the total signals, with the exception of filtered amplitude at ≥50-60 dB in P. koraiensis.【Conclusion】Three different types of UAEs at ≥40-50 dB and ≥50-60 dB classified by amplitude features may be good indicators of the vulnerability of xylem embolism as compared with those at 35-40 dB and ≥ 60 dB. Signal filtration using a mathematical approach did not change the values of P₅₀ at ≥40-50 dB and ≥50-60 dB significantly for each of the two species studied, or for the interspecific difference in embolism vulnerability.