Hyperspectral sensing in the visible through shortwave infrared (VSWIR) portion of the spectrum has been demonstrated to provide significant information on the structural and functional properties of vegetation. The application of these techniques, in the context of leaf-level measurements of VSWIR reflectance, has the potential to revolutionize high-throughput methods to phenotype germplasm that optimizes functional traits related to yield and abiotic stress resilience. Here we focus on a set of breeding trials for warm-season legumes, conducted in both greenhouse and field settings. These trials spanned a set of diverse genotypes providing a range of adaptation to drought and yield potential in the context of semi-arid climate cultivation. At the leaf-level, a large set of spectral reflectance measurements spanning 400-2500 nanometers were made for plants across various growth stages in field experiments that induced severe drought. These reflectance measurements were made alongside gas exchange measurements (A-Ci curves) to provide concurrent values of foliar photosynthetic capacity (Vcmax and Jmax). Here we will discuss the development and performance of partial least squares regression (PLSR) algorithms that relate spectral reflectance to these carbon dioxide exchange parameters that control photosynthetic efficiency.