Long-term daily observations of methane and ammonia from space offer powerful constraints on emissions when ingested into data assimilation systems. In this presentation, I will present global emissions estimates for these two reactive gases capturing seasonal cycles and inter-annual variability. These gases are of interest because recent concentration trends have defied what would be expected from emissions inventory estimates. Methane concentrations surged in 2020-21 for reasons that remain debated in the literature, and I will present evidence that the surge was driven by emissions caused by record inundation of eastern Africa related to an Indian Ocean Dipole anomaly; I will also present detailed analysis of China's increasing methane emissions. While ammonia is not well-mixed, concentrations have been trending upwards in many regions. I will argue that ammonia emissions are increasing and will use a novel machine-learning method to suggest that agricultural sources (principally livestock) are the primary driver but climate variables also are non-negligible. I then use my data-driven results to project the climate-driven ammonia emissions penalty to 2100 and show that my estimates are comparable to results from mechanistic land models.