Long�term trend of topside ionospheric electron density derived from DMSP data during 1995-2017

In recent decades, significant efforts have been made to characterize and understand the global pattern of ionospheric long-term trend. However, little attention has been paid to the topside ionosphere trend. In this study, the unique in situ data measured by series Defense Meteorological Satellite Program (DMSP) satellites were utilized to derive the long-term trend of the topside ionosphere for the first time. We checked carefully data quality, gap, and consistency between different satellites for both electron density and ion temperature, and compared the techniques of artificial neuron network (ANN) and multiple linear regression methods for deriving the trend. The electron density (N-e) trend in the middle and low latitudes at similar to 860 km around 18 MLT was derived using the ANN method from 1995-2017. The trend from DMSP observations has a mean magnitude ranging from similar to- 2% to similar to 2% per decade, with clear seasonal, latitude and longitude variations. The derived trend was evaluated by directly comparing with the simulated trend at 500 km from the NCAR-TIEGCM driven by realistic changes of CO2 level and geomagnetic field. The observed and simulated trends have similar geographic distribution patterns at 18 MLT. The good agreement between the observed trend around 860 km and the simulated trend near 500 km implies that the physical processes controlling the N-e trends above the peak height might be identical. Further control simulations show that the geomagnetic field secular variation is the dominant factor of the electron density trend at around 500 km, rather than the CO2 long-term enhancement.