Biodiversity describes the variety of life from genes to species and ecosystems, providing essential ecosystem services for humans. However, global biodiversity is in peril due to unsustainable human appropriation of ecosystem products and ecosystem transformations to other utilisations. Meanwhile, due to complex spatial patterns and highly dynamic changes, data describing and explaining biodiversity are also heterogeneous. Studies that used longitudinal and cross-sectional data to investigate global diversity change sometimes draw divergent inferences. This suggests an urgent need for helpful diversity estimation to account for the complex interactions amongst species, humans, and natural and built environments. In this thesis, I focus on spatial sample coverage, scale, structure and human-nature interaction that may influence diversity estimation. I also demonstrate that proper diversity estimates are essential for evidence-based conservation by testing whether conservation translocation (i.e., rewilding) can fill native functional gaps effectively. First (Chapter 2), based on neutral community models, I demonstrate that spatial sampling coverage and scales can reconcile divergent inferences on global diversity trends. I show that estimated diversity trends in a landscape with declined diversity can be both increased and decreased in the case where spatial sampling coverage is small, and the local survey scale is imprecise. Next (Chapter 3), I analyse the spatial distribution of diversity time-series recognised as having similar behaviours in temporal variation. I find that spatial diversity patterns can be independent of geographic structure if incorporating information of continued diversity change in the past. Expanding to a perspective of longitudinal analyses (Chapter 4), I illustrate that survey behaviours may respond differently to habitat change during long-term biological survey programmes. I find a significant proportion of survey cessations were associated with habitat change. As habitat change often causes diversity loss, this implies a failure in recording decreased diversity in studies estimating temporal diversity trends. Finally (Chapter 5), I shift to a conservation context and contrast rewilding with three other processes that influence native diversity. I show that effects of different processes vary much in infilling or expanding native functional space and functional diversity change caused by rewilding species are idiosyncratic. Overall, this thesis contributes to elucidating the relationship between actual and estimated nature across scales, notably in diversity estimates. It also demonstrates how spatial structure, scale, and community history may influence estimated diversity patterns and the direct causal implications for shaping conservation strategies. Together, these findings form a foundation for expanding our understanding of the ecological implications of diversity estimation and the mechanisms underpinning this (Chapter 6).