ARTICLE | doi:10.20944/preprints202310.1345.v1
Subject: Environmental And Earth Sciences, Geography Keywords: area studied; BLFR model; BI-LSTM-CRF; improved heuristic disambiguation method; feature template; random forest
Online: 23 October 2023 (05:43:30 CEST)
Geospatial knowledge in massive academic papers can provide knowledge services such as location-based research hotspot analysis, spatio-temporal data aggregation, research results recommendation, etc. However, geospatial knowledge often exists implicitly in literature resources in unstructured form, which is difficult to be directly accessed and mined and utilized for rapid production of massive thematic maps. In this paper, we take the geospatial knowledge of the area studied as an example and introduce its extraction method in detail. An integrated feature template matching and random forest classification algorithm is proposed for accurately identifying research areas from the abstract texts of academic papers and producing thematic maps. Firstly, the precise recognition of geographical names is achieved step by step based on BiLSTM-CRF algorithm and improved heuristic disambiguation method; then, the area studied is extracted by the designed integrated feature recognition template of area studied using random forest classification algorithm, and a fast thematic map is designed for the knowledge of area studied, topic and literature. The experimental results show that the area studied recognition accuracy can reach 97%, the F-value is 96%, and the recall rate reaches 96%, achieving high accuracy and high efficiency of area studied extraction in text. Based on the geospatial knowledge, the thematic map can achieve the effect of fast map formation and accurate expression.
REVIEW | doi:10.20944/preprints202306.2175.v1
Subject: Environmental And Earth Sciences, Other Keywords: afforestation; baseflow; deforestation; FLR; infiltration trade-off; reforestation; soil degradation
Online: 30 June 2023 (07:22:41 CEST)
The large areas being targeted for tropical forest restoration as part of the UN Decade on Ecosystem Restoration will have major consequences for the flow of water through landscapes. Whilst the prevailing mantra that ‘more forest implies less streamflow’ remains true in terms of annual water yields, we demonstrate that opportunities for increased tree cover to improve seasonal flow regimes of streams, particularly baseflows, are important. We discuss several potential positive feedbacks of forest restoration on hydrological processes at various scales, including ‘trade-offs’ between changes in vegetation water use and infiltration after foresting degraded land; the recovery of the capacity of vegetation to capture ‘occult’ precipitation in specific coastal and montane settings; and enhanced moisture recycling and transport at various scales. Modelled changes in baseflow after foresting all degraded land climatically capable of carrying forest in the tropics suggested a positive effect in 10% of the land. For an additional 8%, the effect was predicted to be about neutral (<2 mm/y). We conclude that a more positive narrative regarding the relationship between tropical forestation and water availability is justified. It is time for greater involvement of hydrologists and atmospheric scientists in the development and assessment of forest landscape restoration efforts.