RT journal article T1 Recovery of below-ground associations in restored Brazilian Atlantic Forest A1 Merloti, Luis Fernando A1 Ochoa Hueso, Raúl A1 Zandt, Dina in 't A1 Veen, G. F. A1 Bieluczyk, Wanderlei A1 Rodrigues, Ricardo Ribeiro A1 Mendes, Lucas William A1 Tsai, Siu Mui A1 van der Putten, Wim H. A2 Biología K1 Ecological restoration K1 Active restoration K1 Assisted restoration K1 Soil microbial communities K1 Co-occurrence networks K1 Soil physical-chemical properties AB Soil consists of abiotic and biotic components that sustain biodiversity and forest ecosystem functioning.However, soil restoration has been understudied, and restoration efforts often overlook it. Here, we apply anecological coupling approach using correlation-based network analysis to assess the joint recovery of key abiotic,biotic, and functional soil components during forest restoration in the Brazilian Atlantic Forest. We analyzed soilmicrobial composition as the biotic component, soil physico-chemical characteristics as the abiotic component,and microbial biodiversity, carbon stocks, and greenhouse gases as functional components. Two restorationmethods were studied: (i) active restoration on former sugarcane fields and (ii) assisted restoration on formerpasturelands. We examined chronosequences of early-, intermediate-, and late-stage restored forests, comparingthem to three reference forests. We show that active forest restoration on former sugarcane fields initially disrupted soil associations, but this disruption was overcome as the forest further developed. Active restorationincreased the coupling between soil components and created a co-occurrence network with strong linkagesbetween abiotic and functional soil components. However, the late-stage restored forest remained different fromall three reference forests. Assisted forest restoration on pastures already resulted into coupling during thetransition from pasture to forest, while coupling and network structure of late-stage restored forest was notdifferent from the native conserved reference forest. The observed coupling was the result of strong links betweenbiotic and functional soil components. Both methods facilitated soil recovery, but coupling in assisted restorationwas not different from the native conserved reference forest, while actively restored soils remained distinct fromboth native conserved and degraded reference forests. We propose that actively restored forests may need moretime to converge with reference forests or that active restoration fosters a novel forest soil state. PB Elsevier SN 0929-1393 YR 2025 FD 2025-08-23 LK http://hdl.handle.net/10498/38121 UL http://hdl.handle.net/10498/38121 LA eng DS Repositorio Institucional de la Universidad de Cádiz RD 22-nov-2026