I changed the way the flux J*E^3 is calculated... see write-up, page 4. This seems to result in a lower estimate: