摘要：In the context of sustainable agricultural management,drought monitoring plays a crucial role in assessing the vulnerability ofagriculture to drought occurrence. Drought events are very frequent in theIberian Peninsula (and in Portugal in particular), and an increase infrequency of these extreme events are expected in a very near future.Therefore, the quantitative assessment of the natural-ecosystemvulnerability to drought is still very challenging, mainly due to thedifficulties of having a common definition of vulnerability. Consequently,several methods have been proposed to assess agricultural vulnerability. Inthis work, a principal component analysis (PCA) was performed based on thecomponents which characterize the exposure, sensitivity and adaptivecapacity of the agricultural system to drought events with the aim ofgenerating maps of vulnerability of agriculture to drought in Portugal.Several datasets were used to describe these components, namely droughtindicators, vegetation indices and soil characterization variables. Acomparison between the PCA-based method and a variance method using the sameindicators was performed. Results show that both methods identify Minho andAlentejo as regions of low and extreme vulnerability, respectively. Theresults are very similar between the two methods, with small differences incertain vulnerability classes. However, the PCA method has some advantagesover the variance method, namely the ability to identify the sign of theindicators, not having to use the indicator–component subjectiverelationship, and not needing to calculate weights. Furthermore, the PCA method is fullystatistical and presents results according to prior knowledge of theregion and the data used.
