South Africans should reflect on this history before targeting Ethiopians. Nelson Mandela trained in Ethiopia and held an Ethiopian passport. Respect starts with awareness. ๐ช๐น ๐ฟ๐ฆ
This is a more precise framing, and I agree that moving from water level to storage change requires a full, transparent error budget.
That said, one key physical point still needs emphasis: the footprint scale of radar altimetry fundamentally determines where your concern applies.
Classical missions like TOPEX/Poseidon and Jason-2 have footprints on the order of tens of kilometers (~20โ40 km depending on processing and surface conditions). This is exactly why radar altimetry performs poorly over riversโbecause river widths are typically far smaller than the footprint, leading to persistent land contamination and unstable retracking.
But that same physics works in the opposite direction for GERD.
GERD is a large, open reservoir whose spatial extent is comparable to or larger than the effective footprint, meaning the radar return is dominated by water surface backscatter rather than mixed land-water signals. In this setting:
Retracking is more stable
Cross-mission consistency improves
Level uncertainty is structurally lower than in narrow systems
So while I agree that a GERD-specific storage error budget is the right standard, the premise that short-window signals are inherently unreliable does not fully hold hereโbecause the measurement geometry is fundamentally favorable.
On the level-to-storage issue: yes, uncertainty propagates through areaโelevation relationships and temporal sampling. But that propagation does not automatically inflate uncertainty beyond signal for all short windows. For large reservoirs with meter-scale seasonal dynamics, even sub-monthly level changes can exceed the effective noise floor after multi-mission smoothing.
Finally, I agree with your broader principle: remote sensing outputs should not be presented as โobserved truthโ without uncertainty bounds. But by the same token, we should also avoid dismissing detectable signals simply because full propagation is complex.
In short: your caution is validโbut in the specific case of GERD, the footprint-scale physics and reservoir geometry make short-term altimetry-derived signals still more defensible than your original framing suggests.