The impact of OCO-2 seasonally dependent sampling on carbon flux estimation in the northern tropical Africa

The large annual carbon source over northern tropical Africa (NTA), inferred from satellite CO₂, remains highly debated. Using observing system simulation experiments with Orbiting Carbon Observatory-2 (OCO-2) sampling, we show that seasonally dependent sampling can lead to overestimated annual fluxes. These biases arise when prior flux seasonal cycle differs from the assumed truth. Since OCO-2 provides more observations during the non-growing season, posterior fluxes are more constrained in that period. When prior fluxes underestimate the seasonal amplitude, the posterior carbon sink during the growing season is underestimated, leading to a net positive bias. This effect is supported by real OCO-2 data, where we hypothesize that underestimating fire emissions during non-growing season and weaker seasonality of prior fluxes may contribute to overestimated annual fluxes. Our results highlight the need to improve prior flux estimates and expand observational coverage during the growing season to reduce biases in regional carbon budget assessments over NTA.