Data gaps and trend bias relevant to space climate studies

Long-term variations in the upper atmosphere, and particularly in the ionosphere, are a key topic in space climate studies. Their relevance lies not only in their connection with the increasing concentration of greenhouse gases of anthropogenic origin but also in their potential impact on communications and navigation systems which rely on a precise characterization of the ionosphere. Moreover, thermospheric density changes influence orbital drag and the evolution of space debris. Estimating these long-term variations from experimental data requires extended time series, whose reliability depends on a detailed understanding of station history, instrument changes, and data continuity. A general problem is that ionospheric records, given the length needed to analyze long-term variations, often contain data gaps due to instrumental failures, environmental conditions, or missing measurements. In the case of trend estimation, annual means are commonly used to capture interannual variability. These annual means are calculated from monthly median values archived by data centers, which provide invaluable datasets spanning over 70 years in some cases. Typically, annual means are computed only when at least 8 months of data are available, that is no more than 4 gaps. This study analyses how missing monthly data, ranging from one to 8 gaps per year, affect annual mean values and long-term trend estimates in the F2 layer critical frequency (foF2). The results show that deviations, in addition to increase with missing data number, become substantial for 6 or more, quantitatively supporting the commonly applied requirement of at least 8 months per year, which appears to be a safe choice for reliable estimates.