The spread of invasive Australia native Acacia tree species threatens biodiversity and adversely affecting on vegetative structure and function, including plant community composition, quantity and quality worldwide. It is essential to provide researchers and land managers for biological invasion science and management with accurate information of the distribution of invasive alien species and their dynamics. Remotely sensed data that reveal spatial distribution of the earth’s surface features/objects provide great potential for this purpose. Consistent satellite monitoring of alien invasive plants is often difficult because of lack of sufficient spectral contrast between them and co-occurring plants species. Time series analysis of spectral properties of the species can reveal timing of their variations among adjacent species. This information can improve accuracy of invasive species discrimination and mapping using remote sensing data at large scale. We sought to identify and better understand the optimal time window and key spectral features sufficient to detect invasive Acacia trees in heterogeneous forested landscape in South Africa. We explored one-year (January to December 2018) time series spectral bands and vegetation indices derived from optical Copernicus Sentinel-2 data. The attributes correspond to geographical information of invasive Acacia and native species recorded during a field survey undertaken from 21 February to 25 February 2018 over Kwa-Zulu Natal grasslands landscape, in South Africa. The results showed comparable separability prospects between times series of spectral bands and that of vegetation indices.

Substantial differences between Acacia species and native species were observed from spectral indices and spectral bands which are sensitive to Leaf Area Index, canopy chlorophyll and nitrogen concentrations. The results further revealed spectral differences between Acacia species and co-occurring native vegetation in April (senescence for deciduous plants), June-July (dry season), September (peak flowering period of Acacia spp) and December (leaf green-up) with vegetation indices (overall accuracy > 80 %). While spectral bands showed the beginning of the growing season (November–January) and peak vegetation productivity (February-March) as the optimal seasons or dates for image acquisition for discriminating Acacias from its co-occurring native species (overall accuracy > 80 %). In general, the use of Sentinel-2 time series spectral bands and vegetation indices has increased our understanding of Australian Acacias spectral dynamics, and proved that the sentinel-2 data is useful for characterization and monitoring Acacias over a large scale. Our results and approach could assist in deriving detailed geographic information of the species and assessment of a spread invasive plant species and severity of invasion.

入侵澳大利亚本土相思树的传播树种威胁生物多样性，并对全球的营养结构和功能（包括植物群落组成，数量和质量）产生不利影响。必须为生物入侵科学和管理的研究人员和土地管理人员提供有关外来入侵物种分布及其动态的准确信息。揭示地球表面特征/物体的空间分布的遥感数据为此提供了巨大潜力。由于外来入侵植物与共生植物物种之间缺乏足够的光谱对比，通常很难对它们进行卫星监测。对物种的光谱特性进行时间序列分析可以揭示它们在相邻物种之间变化的时间。该信息可以大规模使用遥感数据提高入侵物种识别和制图的准确性。我们试图确定并更好地理解足以检测侵入性疾病的最佳时间窗口和关键光谱特征在异乎寻常的树木丛生的横向的金合欢结构树在南非。我们探索了从光学哥白尼Sentinel-2数据得出的一年（2018年1月至12月）时间序列光谱带和植被指数。这些属性对应于2018年2月21日至2月25日在南非夸祖鲁纳塔尔（Kwa-Zulu Natal）草原景观上进行的野外调查期间记录的入侵相思和本地物种的地理信息。结果表明，光谱带的时间序列与植被指数的时间序列之间具有可比的可分离性前景。

从对叶面积指数，冠层叶绿素和氮浓度敏感的光谱指数和光谱带中观察到相思树种和天然树种之间的显着差异。结果进一步揭示了相思树种与共同存在的天然植被在4月（落叶植物的衰老），6月至7月（旱季），9月（相思树的高峰开花期）和12月（叶绿化）之间的光谱差异。植被指数（总体准确度> 80％）。光谱带显示出生长季节的开始（11月至1月）和峰值植被生产力（2月至3月）是获取图像的最佳季节或日期，以区分相思树来自其同时出现的本地物种（总体准确度> 80％）。总的来说，使用Sentinel-2时间序列光谱带和植被指数增加了我们对澳大利亚金合欢光谱动力学的理解，并证明了前哨2号数据可用于大规模表征和监测金合欢。我们的结果和方法可以帮助获得该物种的详细地理信息，并评估传播的入侵植物物种和入侵严重程度。