Scots pine management under climate change: a dual-model simulation experiment

SimulationsForest management projections of Pinus sylvestris virtual stands across xeric, mesic, and humid climatic conditions in Catalonia, NE Spain, were conducted under three climate change scenarios (current climate RCP 0.0, RCP 4.5, and RCP 8.5) and three management alternatives (low-intensity/high-frequency thinning LIT, high-intensity/low-frequency thinning HIT, and no management NM). Simulations were run using two forest dynamics models: the process-based GOTILWA+ (Gracia et al., 1999) and the gap model SORTIE-ND (Canham et al., 2005).The simulation outputs were compiled into three files:A raw RDS file containing the complete set of simulation outputs, including all 10 repetitions for each of the 54 simulation cases, capturing stochasticity in stand dynamics.An aggregated dataset containing averaged values across the 10 repetitions, CSVTimber production metrics, calculated for each harvesting interval, CSVThe three datasets include identifiers for model, climatic condition, climate scenario, management alternative, simulation year (or harvest interval), and stand-level attributes such as basal area, DBH, timber production metrics.Fields in sim_meansThe dataset sim_means contains the averaged values across the 10 repetitions for each simulation case. Each row corresponds to a unique combination of simulator, site, management alternative, climate scenario, and simulation year. The fields included are:simulator – Forest dynamics model used in the simulation (GOTILWA+ or SORTIE-ND).site – Climatic condition of the virtual stand (xeric, mesic, humid).mng – Management alternative applied (PS00 = no management, PS08 = low-intensity/high-frequency thinning, PS09 = high-intensity/low-frequency thinning).rcp – Climate scenario (current climate = RCP 0.0, RCP 4.5, RCP 8.5).sim_year – Simulation year.BA – Stand basal area (m² ha⁻¹).STV – Standing timber volume (m³ ha⁻¹).DBH – Mean diameter at breast height (cm).YLD – Yield or timber removals derived from management interventions (m³ ha⁻¹).FR– Fire risk index.Background on the Simulation DesignStand dataPinus sylvestris (Scots pine) is a key species in NE Spain due to its growing stock, carbon sequestration capacity, and harvested volume (Gracia et al., 2004). We simulated its response to current management recommendations (adapted from Pique et al., 2017) under three climate scenarios using GOTILWA+ and SORTIE-ND. Parameter estimates were embedded in GOTILWA+, while SORTIE-ND parameters followed Ameztegui et al. (2015).Stand locations from the Spanish National Forest Inventory were selected to represent humid, mesic, and xeric conditions. To ensure consistent inputs across sites and models, we constructed virtual stands with identical structure, populated with 1450 P. sylvestris trees with DBH between 10–20 cm.Climate change scenariosBaseline (1981–2010) climate data were generated by interpolating daily records from SMC and AEMET using meteoland v1.0.3 R package (De Cáceres et al., 2018). Future conditions were represented by RCP 4.5 and RCP 8.5 projections. GOTILWA+ used daily climate series, whereas SORTIE-ND used monthly temperature and precipitation inputs.Management alternativesManagement objectives followed the ORGEST guidelines (Piqué et al., 2017), prioritizing timber production and fire risk reduction. Strategies combined the shelterwood method with thinning from below (González-Olabarria et al., 2017). The three management alternatives simulated were:PS08 or LIT corresponding to the Low-Intensity / High-Frequency Thinning regimePS09 or HIT corresponding to the High-Intensity / Low-Frequency Thinning regimePS00 or NM corresponding to the No Management alternativeThinning schedules were adaptive, triggered when basal area (BA) and diameter at breast height (DBH) thresholds were reached, while shelterwood stages were implemented at fixed 10-year intervals.