Search

We present the microwave line-by-line model AMSUTRAN, which has been developed at the Met-Office (UK) for over 20 years as part of the EUMETSAT-funded Numerical Weather Prediction Satellite Application Facility (NWP SAF). It produces profiles of layer-to-space transmittances that are representative of the swath of a satellite channel, which are used to train coefficients for the fast radiative transfer model RTTOV (Radiative Transfer for TOVS). At its core are absorption routines based on the Millimeter-wave Propagation Model (MPM), with subsequent modifications to its structure and spectroscopy that have been implemented over time. The most significant of these are: adoption of the Curtis-Godson method of determining the most representative quantities for an atmospheric layer, the complete replacement of all oxygen line parameters, replacement of the air-broadened half-width parameters of the 22.235 and 183.31 GHz water vapour lines, the addition of 35 ozone lines from the HITRAN database, and modifications to the dry continua. The impact of each change is shown in terms of the change in Top Of Atmosphere (TOA) brightness temperature simulated for the 22 channels of the Advanced Technology Microwave Sounder (ATMS) satellite instrument. The biggest effect by far is seen in the method of determining layer quantities, with sensitivities up to many degrees kelvin. To date, developments have focused on the 0–200 GHz range as this is the spectral limit of the current microwave radiometers in-orbit, however, a new generation of instruments heralded by the forthcoming Ice Cloud Imager (ICI) planned for launch in 2022, raises the performance requirements of AMSUTRAN to sub-millimetre frequencies.

To better understand the impacts of climate change, environmental monitoring capabilities must be enhanced by deploying additional and more accurate satellite- and ground-based (including in situ) sensors. In addition, reanalysis of observations collected decades ago but long forgotten can unlock precious information about the recent past. Historical, in situ observations mainly cover densely inhabited areas and frequently traveled routes. In contrast, large selections of early meteorological satellite data, waiting to be exploited today, provide information about remote areas unavailable from any other source. When initially collected, these satellite data posed great challenges to transmission and archiving facilities. As a result, data access was limited to the main teams of scientific investigators associated with the instruments. As archive media have aged, so have the mission scientists and other pioneers of satellite meteorology, who sometimes retired in possession of unique and unpublished information. This paper presents examples of recently recovered satellite data records, including satellite imagery, early infrared hyperspectral soundings, and early microwave humidity soundings. Their value for climate applications today can be realized using methods and techniques that were not yet available when the data were first collected, including efficient and accurate observation simulators and data assimilation into reanalyses. Modern technical infrastructure allows serving entire mission datasets online, enabling easy access and exploration by a broad range of users, including new and old generations of climate scientists.

We present a comparison of a reference and a fast radiative transfer model using numerical weather prediction profiles for the Zeeman-affected high-altitude Special Sensor Microwave Imager/Sounder channels 19–22. We find that the models agree well for channels 21 and 22 compared to the channels' system noise temperatures (1.9 and 1.3 K, respectively) and the expected profile errors at the affected altitudes (estimated to be around 5 K). For channel 22 there is a 0.5 K average difference between the models, with a standard deviation of 0.24 K for the full set of atmospheric profiles. Concerning the same channel, there is 1.2 K on average between the fast model and the sensor measurement, with 1.4 K standard deviation. For channel 21 there is a 0.9 K average difference between the models, with a standard deviation of 0.56 K. Regarding the same channel, there is 1.3 K on average between the fast model and the sensor measurement, with 2.4 K standard deviation. We consider the relatively small model differences as a validation of the fast Zeeman effect scheme for these channels. Both channels 19 and 20 have smaller average differences between the models (at below 0.2 K) and smaller standard deviations (at below 0.4 K) when both models use a two-dimensional magnetic field profile. However, when the reference model is switched to using a full three-dimensional magnetic field profile, the standard deviation to the fast model is increased to almost 2 K due to viewing geometry dependencies, causing up to ±7 K differences near the equator. The average differences between the two models remain small despite changing magnetic field configurations. We are unable to compare channels 19 and 20 to sensor measurements due to limited altitude range of the numerical weather prediction profiles. We recommended that numerical weather prediction software using the fast model takes the available fast Zeeman scheme into account for data assimilation of the affected sensor channels to better constrain the upper atmospheric temperatures.

This paper gives an update of the RTTOV (Radiative Transfer for TOVS) fast radiative transfer model, which is widely used in the satellite retrieval and data assimilation communities. RTTOV is a fast radiative transfer model for simulating top-of-atmosphere radiances from passive visible, infrared and microwave downward-viewing satellite radiometers. In addition to the forward model, it also optionally computes the tangent linear, adjoint and Jacobian matrix providing changes in radiances for profile variable perturbations assuming a linear relationship about a given atmospheric state. This makes it a useful tool for developing physical retrievals from satellite radiances, for direct radiance assimilation in NWP models, for simulating future instruments, and for training or teaching with a graphical user interface. An overview of the RTTOV model is given, highlighting the updates and increased capability of the latest versions, and it gives some examples of its current performance when compared with more accurate line-by-line radiative transfer models and a few selected observations. The improvement over the original version of the model released in 1999 is demonstrated.