An ontological system for interoperable spatial generalisation in biodiversity monitoring

Adapting a Natura 2000 field guideline for a remote sensing-based assessment of heathland conservation status

“…They stated that up to 65% of 'change' could potentially be caused by observer error. This error is even enhanced within NATURA 2000 areas, because local experts tend to overestimate locally relevant species and underestimate locally abundant species (Förster et al, 2008;and further discussed by Nieland et al, 2015). Foody (2008) concludes that the reliability of a map should always be judged within its context in order to reduce inappropriate criticism and false perceptions of remote sensing-based results.…”

“…An accuracy assessment via this matrix should be interpreted as a comparison between two maps (a field-derived and a remote sensing-derived) which both contain a level of uncertainty, rather than a comparison between the remote sensing map and a true reference (Foody, 2008). The later method can either produce smoothing effects by using too coarse resolution or result in intra-class variations if the spatial resolution of the data is very high (Nieland et al, 2015). In this context Hearn et al (2011) reported inconsistencies in repeated vegetation mapping efforts.…”

“…Some parameters for the evaluation, e.g., coverage of open soil or lichens, should rather be considered based on mapping units. Here, object-based analysis (Mücher et al, 2013) or a kernel-based approach (Nieland et al, 2015) could be solutions for finding units of homogeneous vegetation. An alternative could be an adaptation of the guidelines to take account of the capabilities of remote sensing.…”

“…The differentiation of the mainly addressed habitat type European dry heaths (H4030) from similar habitats in the study area (H2030 Dry sand heaths with Calluna and Genista) could be accomplished by adapting the species index and by an additional information layer representing the amount of sand. However, these two habitats are difficult to distinguish via remote sensing and were merged in other studies as well (e.g., Nieland et al, 2015). Moreover, the latter habitat is of minor importance in the study area.…”

Synergetic use of Sentinel‐1 and Sentinel‐2 for assessments of heathland conservation status

“…According to Nieland et al. (), the reasons in mismatching field‐based data and remote sensing data (besides methodological uncertainties and field‐based mapping errors) are given in an improper scale‐match between field data and the resolution of the remotely sensed data (Small ). This can either produce smoothing effects or a high intra‐class variation.…”