Big data for forecasting the impacts of global change on plant communities

摘要: Aim Plant distributions and vegetation dynamics underpin key global phenomena, including biogeochemical cycling, ecosystem productivity terrestrial biodiversity patterns. Aggregated remotely collected ‘big data’ are required to forecast the effects of change on plant communities. We synthesize advances in developing exploiting big data ecology, identify challenges their effective use studies. Location Global. Methods We explored databases, catalogues registries with respect accessibility, geographical taxonomic extent, sample bias other types uncertainty, from perspective both users contributors. identified four kinds needed predict impacts populations communities using spatially explicit models: sensed environmental maps, species occurrence records, community composition (plots) traits, especially demographics. Results Digital data, most mature class discussed herein whereby protocols for archiving, discovering analysing them have developed over three decades. Species locality records being aggregated into databases that easy search access, while methods addressing uncertainties a major research focus, better spatial representation is still needed. Plot inventories tremendous potential monitoring modelling but tend be restricted forests or concentrated certain areas. Ongoing efforts aggregate plot trait multiple sources challenged by heterogeneous coverage, attributes lack standards. Main conclusions Future goals include systematic frameworks selecting geospatial improving tools assessing quality increased aggregation discoverability data. scientists, not sensors, provide more meaningful insights when collectors involved analysis.