L-band InSAR snow water equivalent retrieval uncertainty increases with forest cover fraction

Abstract

There is a pressing need for global monitoring of snow water equivalent (SWE) at high spatiotemporal resolution, and L-band (1-2 GHz) interferometric synthetic aperture radar (InSAR) holds promise. However, the technique has not seen extensive evaluation in forests. We evaluated this technique across varying forest canopy conditions using eight InSAR pairs collected at the Fraser Experimental Forest, Colorado, USA by NASA UAVSAR during the 10 week NASA SnowEx 2021 Campaign. Compared with in situ measurements, we found root mean squared errors (RMSEs) of 14-17 mm for SWE changes in forest cover fractions (FCF) < 0.40, but RMSEs increased to 33-40 mm at FCF > 0.50. Statistical distributions between normalized lidar snow depths and normalized UAVSAR SWE were similar at FCF < 0.5, but diverged at FCF > 0.50. Thus, the upcoming NISAR L-band satellite has strong potential for global snowpack monitoring, including below sparse to moderate forest cover.

Citation

Bonnell, R.; Elder, K.; McGrath, D.; Marshall, H. P.; Starr, B.; Adebisi, N.; Palomaki, R. T.; Hoppinen, Z. 2024. L-band InSAR snow water equivalent retrieval uncertainty increases with forest cover fraction. Geophysical Research Letters. 51: e2024GL111708.