Modeled material behavior of sea ice assumes some form of elastic/viscous/plastic rheology. The objective of this work is to test this assumption against observed measurements over annual cycles at scales commensurate with climate models (100 km). Using an archive of stress buoys with GPS positioning from Beaufort Sea ice campaigns in 1998 (SHEBA) and 2001 (Beaufort01), we compare observed stresses between 2 years and buoy locations from SHEBA in the Los Alamos CICE model version 3. Key differences between modeled and observed stress at these scales primarily arise from their inherent differences in design. A numerical model seeks to define the yield surface (or breaking threshold) of sea ice over a specified numerical footprint. Conversely, time series of in situ measurements are collected from point source buoys placed in the interior of individual ice floes. By design, it is highly unlikely for a stress sensor buoy to be situated in a damage zone where failure is occurring. Hence in principal axis space, stress measurements from buoys most often record stress states within a yield surface, rather than along a yield surface. By analyzing the extreme states in measured stress and corresponding time evolution, we effectively compare fundamental differences between modeled and observed stress. To demonstrate, we show the important process of annealing through cyclic build-up and release of stress with observations showing a symmetric process while the model evolves asymmetrically. We hypothesize that correct reproduction of this process should improve ice-volume estimates in models. Careful tracking of changes in thickness and concentration are shown to be critical. Results shown here encourage further investigation of model sensitivity relative to observed stresses from buoy arrays. While the shape of a yield curve will be extremely difficult to map from observed buoy arrays, understanding of material behavior through process studies provides a more integrative approach. Use of process studies advances understanding of the material properties of sea ice and more importantly the interplay of material behavior with sea-ice dynamics, thermodynamics and mass-balance considerations. The co-location of stress buoy arrays with ice mass-balance (IMB) buoys and local changes in ice concentration clearly advances integrated analysis capabilities for such process studies. Discussions and conclusions examine the design of integrated measurements such as these through efforts supporting an Arctic Observing Network.