Abstract: To uncover the spectral characteristics of the main soils in Songnen Plain, northeastern China, and the spectral fast testing methods for soil parameters, the room hyperspectral reflectance of soil samples from Nongan county, Songnen Plain was measured. With the help of statistic analysis methods, the relationship between soil parameters and reflectance was analyzed and spectral predicting models for soil organic matter content were also achieved. The results are as follows: ⑴ organic matter is the main factor impacting on soil reflectance characteristics of Songnen Plain, and because of the differences of soil parent matter and mechanical composition, even the same type soils from different area have different spectral characteristics; ⑵ the impact of Fe on soil reflectance in Songnen Plain is not apparent; ⑶ organic matter predicting models basing on spectral properties can be used for fast testing; ⑷ the correlation between organic matter and total N is significant, so the overall total N status can be discovered with organic matter spectral predicting models.

Key words: soil, spectrum, organic matter, Fe, Songnen Plain