Assessment of Equatorial Plasma Bubble Impacts on Ground-Based Augmentation Systems in the Brazilian Region

Abstract

A preliminary Brazilian ionospheric anomaly threat model for Ground Based Augmentation Systems (GBASs) was developed from a comprehensive analysis using many days of ionospheric data from Brazil to reflect low-latitude conditions. Most anomalous ionospheric spatial gradients in this model were caused by nighttime Equatorial Plasma Bubbles (EPBs). In particular, the largest observed EPB-induced gradient (similar to 860 mm/km) is two times larger than the upper bound (425 mm/km) on spatial gradients for the Conterminous U. S. (CONUS). The higher bound in the Brazilian ionospheric threat model has a significant effect on GBAS performance and availability. More detailed performance evaluation based on the Brazilian threat model is needed before GBAS can be made operational there.

This paper first assesses the performance of GBAS with Position Domain Geometry Screening (PDGS), which has been used in CONUS. Using the parameters of the preliminary EPB threat model, sets of worst-case geometries among an EPB gradient front and satellite Ionospheric Pierce Points (IPP) are generated. To calculate undetected range errors for hypothetical GBAS users, we derive an analytical ionosphere-induced differential range error model. Based on these inputs, we analyze the performance of GBAS under worst-case ionospheric conditions. The results show that inflating the standard deviation of the vertical ionospheric gradient only is not sufficient to completely remove potentially unsafe satellite geometries. This leads to additional inflation of the standard deviation of pseudorange correction errors. If the maximum EPB-induced gradient is larger than about 600 mm/km, the resulting CATegoryI (CAT-I) GBAS availability is significantly degraded, making it difficult to meet the system requirement using this approach.

For this reason, we have also employed Monte Carlo analysis. The key difference from the previous approach is that credit is taken for a prior probability of an extreme EPB event instead of assuming that worst-case storms occur with a probability of one in PDGS. This stochastic approach assesses the overall user impact by running many anomalous ionospheric trials based upon numerous combinations of threat model parameters. For each subset geometry, it is assumed that either one or two satellites are affected by the worst-case gradient while the remaining satellites are affected by anomalous but nonworst- gradients. Therefore, the worst-case impact should be approximated by inclusion in the distribution of simulation results. A real-time mitigation scheme based on the Monte Carlo method, which also requires sigma parameter inflation, is proposed that takes credit for the prior probability. System availability evaluation using the Monte Carlo approach with a single worst-case satellite impact shows that the availability penalty is significantly reduced compared to the previous (worst-case) method.