A plot sampling strategy for estimating the area of olive tree crops and olive tree abundance in a Mediterranean environment

Olive tree (Olea europaea L.) is a long-living evergreen plant with relevant cultural, economic and landscape relevance. Accurate inventory and mapping of olive surface and olive abundance represents a central issue to support the olive production system. With reference to the cultivation, there is a high heterogeneity and complexity in the cultivation of olive trees. In addition, olive is cultivated in a wide variety of soils and can tolerate a broad range of physicochemical conditions. Such heterogeneity in production systems is reflected in the large variability in olive surfaces, which poses some challenge in accurately estimating the area of olive tree crops and olive tree abundance via traditional inventory approaches, as e.g., commonly adopted for national forest inventory. From a methodological point of view, the complexity and heterogeneity of olive tree crops can be comparable to the problem of accurately estimating tree outside forests (TOF) attributes. TOFs are defined as small groups of trees within rural and urbanized areas not classified as forests. Only relatively recently the importance of including TOF to integrate forest inventory estimates have been addressed by some studies Similarly to traditional forest inventory, TOF inventory strategies are usually based on environmental sampling scheme such as plot sampling, and its integration with information derived from remotely sensed imagery. A successful plot sampling strategy (PLIS50) have been recently verified in TOF from Fattorini et al. (2016). In this study we verified whether the plot sampling approach proposed for TOF is suitable for estimating the area of olive tree crops and olive tree abundance. For the purpose, the approach was tested in a case study where the census of the olive crop area, number of olive crops (polygons) was manually conducted from photo-interpretation of high resolution aerial orthoimagery, which were used as benchmark to test the effectiveness of the point sampling approach. The main results of this study is that the plot sampling method, formerly applied in TOF inventorying , can be applied for estimating olive tree attributes. In our case study we obtained accurate estimates of olive tree attributes, whose RSEs were always below 10%, with a limited sampling of about 6% of the studied population (olive grove polygons), with considerable reduction in cost and time efforts required compared with manual census from photointerpretation. Using robust statistical procedures among countries, should allow obtaining harmonized and comparable information, which can increase the knowledge of olive geographical distribution and structure at its relevant Mediterranean scale.