Stratospheric atmosphere is jointly affected by dynamical， radiative and chemical processes.Since the 1970s， the greenhouse gas emission and ozone depletion caused by anthropogenic activities have driven significant stratospheric cooling， so the tropospheric warming and the stratospheric cooling are both important criteria for global climate change.Observational and modeling studies disclosed that the formation of polar stratospheric clouds at very low stratospheric temperatures is an important condition for rapid ozone depletion.The weak polar vortex event related to the stratospheric sudden warming can even propagate downward to the troposphere and induce the tropospheric circulation anomaly， further leading to the extreme cold event in the middle and high latitude continents.Therefore， it is necessary to clearly understand the causes of stratospheric temperature changes at the background of global warming.Under quasigeostrophic scaling， the eddy heat flux approximation theory within the transformed Eulerian-mean framework is usually used to calculate the dynamical and radiative heating contributions of the Arctic stratospheric temperature changes， but this may bring certain errors.In this paper， according to a new area-weighted averaged thermodynamic equation by Liu and Fu （2019） within the Eulerian-mean framework， a monthly temperature change equation is constructed by time-sliding accumulation.Then using the European Centre for Medium-range Weather Forecasts fifth reanalysis data （ERA5） during 1980-2019， we calculate the climatological monthly temperature increment， dynamical heating， diabatic heating and convective heating term （W term） in the Arctic lower stratosphere， respectively， during the whole period 1980-2019 and two sub periods （1980-1999 and 2000-2019）， and discuss the best reference latitude for the eddy heat flux approximation theory.The results show that the newly derived W term reduces the imbalance between the Arctic temperature increment term and the sum of cumulative dynamical and diabatic heating by half in winter and spring.At 100 hPa， W term changes with month and latitude， nontrivial especially in winter and spring.Therefore， the best reference latitude for the eddy heat flux approximation should be near the zero contour line of the W term.Further verification shows that the reference latitude can be taken at 50°N.