Measuring plant diversity through field observation is a resource consuming process, so focusing it on hotspots or sinks would increase efficiency. To estimate this priori diversity distribution, remotely sensed essential biodiversity variables, obtained from high spatial and spectral resolution satellite imagery, are gaining attention. Indicators based on spectral diversity are rapidly developing, but their applicability needs to be tested across different ecosystems and conditions. Therefore, we compared field observed alpha plant diversity indices of temperate forest on Mount Usu with remotely sensed spectral diversity indicators. As both scale (resolution) and inclusion of surrounding pixels is known to influence the results of spectral diversity analysis, we used two scales and two spectral plot sizes. The small scale analysis was performed using 0.82 m resolution, while the large scale one was carried out on 3.28 m resolution Ikonos imagery. Pixel values were obtained from square (10 m sides) and circular (10 m radius) spectral plots for all locations. Regression analysis showed weak or no correlation between field and spectral based indices, with no clear effects of scale and spectral plot size. Our findings are consistent with those of other studies which found significant but weak correlation between field and spectral indices.