Modelling propagation impairments of Earth-Space links using Numerical Weather Prediction tools
(2020)
Files
thesis_Laurent_Quibus.pdf
Open Access - Adobe PDF
- 10.26 MB
Details
- Authors
- Supervisors
- Vanhoenacker-Janvier, Danielle
- Abstract
- Communication channels between the Earth and satellites are in the on-going process to be scaled up to higher and higher carrier frequencies above 20 GHz. The change is driven forward as an increasing number of services fill up the spectral bands allocated at lower frequencies, and as using higher frequencies also results in larger bandwidths, antenna gains and directivity. However, challenges to high frequency transmissions are posed by the stronger impairments of the signal as it propagates through the troposphere. In particular, the power of the signal is attenuated due to the presence of oxygen, water vapour, clouds, rain, and turbulence. The attenuation affects the signal-to-noise ratio (SNR) by up to tens of dB so that Fade Mitigation Techniques (FMTs) become needed. Ideally, the design of FMTs should be based on direct measurements of the attenuation. Earth-space propagation experimental campaigns offer this reference thanks to spaceborne beacons emitting continuously at selected frequencies. These experiments are however costly, limited to a few receiving stations in a certain region, and lasting typically only a few years. The stations also require additional equipments such as radiometers in order to obtain the total attenuation and not just the contribution from rain events and amplitude scintillation. There is therefore an interest for models capable to derive the attenuation components from alternative measurements, so as to supplement the beacon data or fill the gaps in their coverage. Numerical Weather Predictions (NWPs) are important candidates as they aim to re-create, or even forecast, the state of the atmosphere (pressure, temperature, water, ...) starting with the assimilation of various measurements collected globally. This thesis principal objective is hence the investigation of the performances of a NWP-based simulator of the attenuation within the context of propagation experiments. The thesis opens by introducing further the problems in Earth-space propagation, FMTs, propagation impairments, propagation experiments, and NWPs. A short review of the propagation experiments shows their current limits in frequency (1 to 3, ~ 20 GHz and < 50 GHz), coverage (mostly temperate), duration (usually ~ 1 − 5 years), number of stations (from 1 to rarely ~ 20), and orbits (overwhelmingly geosynchronous orbits (GSO)). A short review of the previous usage of NWPs demonstrates the rarity of results simulated over the long-term (> 1 year), with high resolutions (< 5km horizontally, < 1 h), or for non-GSO links. It also points out the existence of two methods for the rain attenuation (rain volumetric content or rain rate), and of two methods for the cloud attenuation (NWP parametrisation or Salonen/Mattioli). The thesis continues on the description of the propagation models, NWP models, interpolations or conversion of coordinates, and error metrics involved in the simulation of the attenuation from NWP data and its comparison with other reference datasets (beacon, radiometer, and weather radar). The theoretical foundations of the work introduce few novelties, with many of the propagation models following the recommendations from the International Telecommunication Union-Radiocommunication (ITU-R), and with the implementation of the simulator close to the existing solutions in the literature, especially the work of Jeannin et al.. This thesis aims however to test competing approaches, including the two alternatives for rain and cloud attenuations, or different NWP parametrisations. Some practical innovations are presented such as adaptive NWP time steps, and a post-processing without non-linear interpolations. Other originalities concern the estimation of the radiometric accuracy, or the computation non-GSO statistics either directly or from GSO statistics. The novelty is stronger for the validation of the performances in two phases. In the first validation phase, long-term simulations (5 years) are performed for GSO and non-GSO links (19.701, 26, 39.402 and 75 GHz, Alphasat and METOP) at Louvain-la-Neuve (BE), at resolutions of 4km and 5 min, with the Weather Research and Forecasting (WRF) model. These simulations are compared to equivalent ones using the ALARO model (4km and 1 h), to weather radar data for rain (5 min), and to the ITU-R references. Among the results, the correlations in attenuation (at 30° in elevation) between WRF and ALARO are ~ 0.91 for oxygen, ~ 0.78 for water vapour, but only ~ 0.11 for clouds, and even lower for rain. The agreements in complementary cumulative distribution functions (CCDFs) are acceptable for the gases. For the clouds, there is an underestimation by the NWPs (by ~ half) with respect to the ITU-R, though the Salonen/Mattioli models overestimate the ITU-R at high probabilities. Clouds also display a marked instantaneous dependence in azimuth and elevation, but a weak inter-annual variability. Rain attenuation has a strong inter-annual variability, but starts to stabilise after 5 years. The NWPs underestimate the ITU-R CCDFs for rain, though with the WRF rain rate it is by only ~ 1 dB and almost matching the radar results. In the case of non-GSO CCDFs, different attenuations are predicted from non-GSO time series than from a combination of GSO CCDFs, which suggests a design based on the latter is suboptimal. In the second validation phase, the nonrainy attenuation (gases and cloud) is assessed thanks to four months of Alphasat beacon and radiometric data at Spino d’Adda (IT). Correlations and root-mean square errors between WRF (Tiedtke) and the radiometer are: 0.88 and 0.11 dB at 19.701 GHz, 0.59 and 0.26 dB at 39.402 GHz. The errors are comparable to the limits in radiometric(< 0.1 dB) and beacon processing (~ 0.2 to 0.5 dB) accuracies. The simulator could then act as a radiometer replacement within a propagation experiment. After the validation, the thesis presents other applications of NWP-derived propagation impairments. Simulated Alphasat gaseous and cloud attenuation CCDFs during one year are shown for nine European ground stations. Other non-GSO CCDFs with the METOP satellites are shown for Graz (AT). The observations made from all those CCDFs are in line with the results obtained during the validation. Finally, an attempt to relate strong scintillation conditions with NWP data is described for a low elevation link in Isfjord (NO). At the end, the performances are seen to be: good for the gases, sufficient as a radiometer’s substitute for the clouds, acceptable statistically over multiple years for rain, only qualitative but promising for scintillation, and suggesting the ITU-R scaling from GSO to non-GSO CCDFs is suboptimal. Many perspectives remain open, e.g. at even higher radio-, or optical, frequencies.
- Affiliations
Citations
Quibus, L. (2020). Modelling propagation impairments of Earth-Space links using Numerical Weather Prediction tools. https://hdl.handle.net/2078.5/167820