岷江上游亚高山典型森林植被高光谱特征识别

在高光谱数据分类应用中,地物光谱特征分析是对地物进行分类和检索的基础性工作。选取禾本科斑竹、草本科蕨类、荨麻科冷水花、杉科杉木和棕榈科棕榈树等5种岷江上游亚高山森林植被进行实地光谱测量,建立高光谱相似性度量参量,如欧式距离(Euclidean distance,ED)、光谱角度(spectral angle mapper,SAM)、光谱信息散度(spectral information divergence,SID)、SID和SAM混合SID(TAN)以及基于道格拉斯-普克算法的光谱降维距离(spectral distance based on Douglas-Peucker,SDDP)度量算法,定量分析对亚高山森林植被的识别能力。研究结果表明:5种亚高山森林植被光谱特征的差异主要表现在光谱曲线反射波峰和波谷位置;ED对冷水花的相对光谱识别概率最高;SID和SID(TAN)对斑竹与蕨类的识别概率最高;SDDP对杉木的识别概率最高;SAM,SDDP,ED,SID(TAN)和SID这5种光谱相似性测度算法对亚高山森林植被的相对光谱识别熵分别是1.51,1.59,1.61,2.16和2.18,说明光谱角度制图具有较高的识别能力;而道格拉斯-普克光谱检索算法是在提取光谱曲线特征向量的基础上进行相似性测度,其降低了光谱检索的时间频率,在保证相近识别能力的条件下,能够大大提高程序的检索效率,是一种快速有效的高光谱特征匹配和检索算子。

Vegetation-based bioindication of humus forms in coniferous mountain forests

Humus forms, especially the occurrence and the thickness of the horizon of humified residues （OH）, provide valuable information on site conditions. In mountain forest soils, humus forms show a high spatial variability and data on their spatial patterns is often scarce. Our aim was to test the applicability of various vegetation features as proxy for OH thickness. Subalpine coniferous forests dominated by Picea abies （L.） H. Karst. and Larix decidua Mill. were studied in the Province of Trento, Italian Alps, between ca. 900 and 22o0 m a.s.1. Braun-Blanquet vegetation relevds and OH thickness were recorded at 152 plots. The vegetation parameters, tested for their suitability as indicators of OH thickness, encompassed mean Landolt indicator values of the herb layer （both unweighted and cover-weighted means） as well as parameters of vegetation structure （cover values of plant species groups） calculated from the releves. To our knowledge, the predictive power of Landolt indicator values （LIVs） for humus forms had not been tested before. Correlations between OHthickness and mean LIVs were strongest for the soil reaction value, but indicator values for humus, nutrients, temperature and light were also significantly correlated with OH thickness. Generally, weighting with species cover reduced the indicator quality of mean LIVs for OH thickness. The strongest relationships between OH thickness and vegetation structure existed in the following indicators： the cover of forbs （excluding graminoids and ferns） and the cover of Erieaeeae in the herb layer. Regression models predicting OH thickness based on vegetation structure had almost as much predictive power as models based on LIVs. We conclude that LIVs analysis can produce fairly reliable information regarding the thickness of the OH horizon and, thus, the humus form. If no releve data are readily available, a field estimation of the cover values of certain easily distinguishable herb layer species groups is much faster than a vegetation survey with consecutive indicator value analysis, and might be a feasible way of quickly indicating the humus form.