Gross primary production (GPP) is a key component of the forest carbon cycle. However, our knowledge of GPP at the stand scale remains uncertain because estimates derived from eddy covariance (EC) rely on semi-empirical modeling and the assumptions of the EC technique are sometimes not fully met. We propose using the sap flux/isotope method as an alternative way to estimate canopy GPP, termed GPPiso/SF, at the stand scale and at daily resolution. It is based on canopy conductance inferred from sap flux and intrinsic water-use efficiency estimated from the stable carbon isotope composition of phloem contents. The GPPiso/SF estimate was further corrected for seasonal variations in photosynthetic capacity and mesophyll conductance. We compared our estimate of GPPiso/SF to the GPP derived from PRELES, a model parameterised with EC data. The comparisons were performed in a highly instrumented, boreal Scots pine forest in northern Sweden, including a nitrogen fertilised and a reference plot. The resulting annual and daily GPPiso/SF estimates agreed well with PRELES, in the fertilised plot and the reference plot. We discuss the GPPiso/SF method as an alternative which can be widely applied without terrain restrictions, where the assumptions of EC are not met.

Gross primary production (GPP) is a key component of the forest carbon cycle. However, our knowledge of GPP at the stand scale remains uncertain because estimates derived from eddy covariance (EC) and EC itself rely on semi-empirical modeling and the assumptions of the EC technique are sometimes not fully met. We propose using the sap flux/isotope method as an alternative way to estimate canopy GPP, termed GPPiso/SF, at the stand scale and at daily resolution. It is based on canopy conductance inferred from sap flux and intrinsic water-use efficiency estimated from the stable carbon isotope composition of phloem contents. The GPPiso/SF estimate was further corrected for seasonal variations in photosynthetic capacity and mesophyll conductance. We compared our estimate of GPPiso/SF to the GPP derived from PRELES, a model parameterised with EC data. The comparisons were performed in a highly instrumented, boreal Scots pine forest in northern Sweden, including a nitrogen fertilised and a reference plot. The resulting annual and daily GPPiso/SF estimates agreed well with PRELES, in the fertilised plot and the reference plot. We discuss the GPPiso/SF method as an alternative which can be widely applied without terrain restrictions, where the assumptions of EC are not met.