UC Berkeley postdoc presents low‑cost ET sensor that matched eddy covariance within 2–10% at wetlands

Carlos Juan, a postdoctoral researcher at UC Berkeley, told the DMEC that a variance‑bone‑ratio (VBR) approach can produce cost‑effective, near‑real‑time evapotranspiration (ET) estimates and could help fill local data gaps in the Delta. "From these testings, we saw a few limitations," he said, but across three wetlands the VBR totals "tracked surprisingly close" to eddy covariance, on the order of "within, I think, 2 to 10%."

What he presented: Juan explained VBR math briefly and described test sites in the Delta region (a Suisun/Hillsdale marsh site, an Oakley/Gilbert Tract site, and Twitchell Island East End). The VBR sensor package combines net radiometers, ground heat flux plates and temperature/relative humidity probes; the processing code is open source. He reported stronger agreement at two open‑water/vegetation sites and larger errors at a site with accumulated dead biomass.

Limitations and next steps: Juan cautioned that VBR is sensitive to the accuracy of net available energy measurements and underperforms under strong advection — a condition common in agricultural lands — and that negative bone‑ratio values must currently be forced positive in his implementation. He proposed a dense observational 'Vortex' network of turnkey VBR stations to produce pixel‑scale ET alongside high‑accuracy reference sites for bias correction and downscaling.

Why it matters: a lower‑cost sensor network could make dense ground observations more affordable, provide near‑real‑time ET estimates for managers, and complement satellite products for local water‑balance and ACP work. Juan invited collaboration on expanding VBR tests to rice paddies and other crop types.

Next steps: the presenter requested guidance on practitioner needs (daily vs. weekly vs. monthly requirements) and plans to expand accuracy assessments across agriculture, forest, wetland and urban gradients.