NRLMSISE-00: A New Empirical Model of the Atmosphere

J.M. Picone, D.P. Drob, and R.R. Meier
Space Science Division

A.E. Hedin
Universities Space Research Association

Introduction: NRL has completed the new NRLMSISE-00 empirical model of the atmosphere for worldwide distribution to operational users and scientists.1 MSIS stands for Mass Spectrometer and Incoherent Scatter Radar, the two primary data sources underlying early versions of the model, and E indicates that the model extends from the ground to space, as opposed to early versions that covered only the upper atmosphere or "thermosphere" (altitude > 90 km). NRLMSISE-00 represents the culmination of an effort to preserve and radically extend NASA's MSIS technology so that future military and scientific users could exploit the model's advantages. The model calculates composition, temperature, and total mass density, and is the standard for international space research. Improvements have focused on the thermosphere, which offers the potential for a number of vital operational and scientific applications.

NRLMSISE-00 accounts for the main drivers of the upper atmosphere: the solar extreme ultraviolet (EUV) flux and geomagnetic heating. The 10.7-cm solar radio flux (F10.7) is the standard proxy for the solar EUV, while the Ap daily geomagnetic index measures the geomagnetic component of space weather. The next section outlines improvements that make NRLMSISE-00 a strong candidate to replace the 30-year-old Jacchia-70 model as the standard for space object orbit determination and prediction by the Navy and the Air Force. The next-generation Air Force orbit model has already adopted the MSIS representation of the lower atmosphere to eliminate a glaring deficiency of Jacchia-70, which works only for altitudes above
90 km.

Revolutionary Improvements: For the first time, this MSIS-class model assimilates total mass density values determined from drag on satellites and other space objects while retaining the traditional mass spectrometer and radar databases. The addition of drag data to the NRLMSIS database on composition and temperature gives the model a foundation superior to that of the Jacchia models, which are based primarily on orbital drag data produced in the 1960s. In contrast, the NRLMSIS database now covers the last four decades, with notable NRL upgrades of the temperature and molecular oxygen (O2) data sets. As a result, the NRLMSISE-00 model has potential applications in precision orbit determination, space object re-entry, ionospheric forecasting, ionospheric D-region absorption of high-frequency signals, and infrasound site location.

Our development work has opened several important areas of scientific study, including the resolution of contradictory measurements of molecular oxygen in the lower thermosphere, enhancement of molecular ions in the ionospheric F-region by geomagnetic storms, and spacecraft drag due to ionospheric oxygen ions and hot atomic oxygen. Molecular oxygen is of critical importance in prediction of the ionospheric F-region and in the inversion of new remote sensing data from the NRL/Space Test Program ARGOS satellite mission. Earlier MSIS-class models lacked data for O2 at higher solar activity and depended primarily on data from mass spectrometers flown by NASA during the 1970s when the EUV flux was low. The early models then used temperature data to extrapolate the estimated O2 concentration to elevated solar conditions.

NRL has now acquired O2 data from solar ultraviolet absorption measurements aboard the NASA Solar Maximum Mission (SMM) to generate NRLMSISE-00 across a wide range of solar EUV and altitude. This has drastically changed the model predictions (Fig. 6). Figure 6(a) shows the logarithm of the ratio of the O2 data to the NRLMSISE-00 model density as a function of solar activity (F10.7). The horizontal line at a value of 0.0 corresponds to the new model (ratio = 1). The vertical bars signify the range of values of the data relative to the model, and the clustering of the data around the NRLMSISE-00 line signifies a good fit across several decades of data, including the SMM data (symbol "K" on the plots). Figure 6(b) shows a similar comparison to the previous model, NASA's MSISE-90, which does not match the weak solar EUV dependence of SMM. Early applications of the new model by ionospheric scientists have shown great promise in resolving past paradoxes and improving the model predictions.

Recent investigations of low Earth orbit (LEO) drag have shown that the operational atmospheric model, Jacchia-70, erroneously attributes nonthermospheric drag sources to atomic helium (He). We now have evidence that these sources consist of "hot" atomic oxygen and ionospheric atomic oxygen ions (O+), which can be of primary importance during the summer at high latitudes and altitudes above 600 km. Since neither of these species is in thermal equilibrium with the thermosphere, the new NRLMSISE-00 model treats them as a new component to drag called "anomalous oxygen." In addition to capturing the effects associated with anomalous oxygen, NRLMSISE-00 provides a superior fit to the corresponding winter Jacchia drag data at altitudes above 600 km, as compared to the Jacchia-70 model. The new model also eliminates deficiencies found in the Jacchia model for the summer hemisphere at very low solar activity.

Fig 6a
Natural logarithm of lower thermospheric [O2] vs 81-day mean F10.7, averaged within bins of 10 flux units. The plot shows the bin-averaged data values normalized by NRLMSISE-00. Vertical bars correspond to the 1s range of normalized [O2] values within each bin. NRLMSISE-00 corresponds to the horizontal line at 0.0. Symbols: C, G, J, +, mass spectrometer data; A, B, D, E, F, M, V, W, solar ultraviolet absorption data (100-150 km); K, SMM data (140-200 km).

Fig 6b
Same as (a), but with data normalized to the older MSISE-90 model, which corresponds to the horizontal line at 0.0.

New Directions: NRL has now become an international center for assimilative upper atmosphere models and is developing new applications of the NRLMSISE-00 model, including analysis of ultraviolet remote sensing data for precise orbit determination and prediction, accurate upper atmospheric composition for ionospheric forecasting, and ground-to-space atmospheric data assimilation for infrasound site location. NRL scientists are also using NRLMSISE-00 along with long-term orbital drag data to evaluate human-induced change in the upper atmosphere over several decades and to test new solar EUV proxies for more accurate prediction of the thermospheric state.

[Sponsored by ONR]


1 J.M. Picone, A.E. Hedin, D.P. Drob, and A.C. Aikin, "NRL-MSISE-00 Empirical Model of the Atmosphere: Statistical Comparisons and Scientific Issues," J. Geophys. Res., doi:10.1029/2002JA009430, in press (2003).