Article Type: Original Article

The goals of this paper are to assess feasibility of using total concentration (TPC) of phytoliths in modern soils and morphotypes in soil assemblages to infer paleovegetation cover and composition. We sampled from undisturbed grassland sites with different values of shrub and grass cover in the Columbia Basin, interior Pacific Northwest, USA. Phytolith concentration in soils were measured and all morphotypes counted. Estimates of vegetation cover were obtained using NDVI values derived from multispectral digital 3 3 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 imagery. Linear regression was used to predict NDVI values based on TPC in soils. Linear regression revealed a direct positive relation between vegetation cover and TPC (R2 = 0.49). Estimates of vegetation cover averaged over multiple years may increase it’s relation with TPC. CCA was used to assess the degree of similarity between field plots. The morphotypes-species correlations for the axes are moderately high (r=0.863 and r=0.660 for the first two axes, respectively), suggesting that phytoliths reflect vegetation plot compositions with moderately high certainty. We conclude that phytoliths reflect modern vegetation in grasslands and shrublands on the Columbia Basin. Estimates of vegetation cover reflect annual and inter-annual trends, while the phytolith assemblages from modern soils reflect the vegetation averaged over a decadal to centennial scale.