Urban local climate zoning simulation coupled with SSP-RCP scenarios:A case study of Nanjing
Land use and cover change(LUCC)is a remarkable manifestation of the impacts exerted by human activities on the natural environment,and land use scenario simulation is one of the main ways to study LUCC.Scenario-based land use change simulation can show multiple possibilities of future cities and is currently a hot spot in the field of land cover change.However,the existing surface classification system of construction land in this filed cannot accurately portray the internal spatial morphology of the city,and the non-uniform simulation scenarios will affect the comparability of the land simulation results.To address these issues,with the help of three typical SSP-RCP coupling scenarios proposed by the international Coupled Model Intercomparison Project and the local climate zones(LCZ)that can depict the internal morphology of the city,this paper carried out spatial and temporal simulation of the LCZ for the near-term(2030),medium-term(2060),and long-term(2090)in the case of Nanjing.The study employed Markov,multiple objective programming(MOP),and future land use simulation(FLUS)models to simulate the spatial and temporal changes of LCZs under different SSP-RCPs coupling scenarios.In terms of quantity prediction of land types,based on the LCZ classification results in the base period,we first used Markov to determine the quantity scale of each LCZ type in different future periods along the historical scenario.Then,structure optimization was carried out by MOP to predict the land demand of LCZ types under different SSP-RCP coupling scenarios.In terms of spatial simulation of land types,20 driving factors such as physical geography,socio-economic,and location conditions were collected,and spatial distribution simulation of the quantity scales of different LCZ types under the future scenarios obtained above was realized through the FLUS model.Finally,the Receiver Operating Characteristic curve corresponding to the probability of suitability of each LCZ type was calculated based on the logistic regression model,and the area under the curve(AUC)value of this curve indirectly reflects the simulation accuracy.Our results showed that:①With respect to the validation of the simulation accuracy,the average AUC value of suitability probability of all LCZ types was 0.73,with relatively high AUC values(＞0.85)for land types of dense high-rise(LCZ 1),compact mid-rise(LCZ 2),and open mid-rise(LCZ 5),while the AUC values of compact low-rise(LCZ 3),sparsely built(LCZ 9),dense trees(LCZ A),and water(LCZ G)ranged between 0.65 and 0.75.②From the LCZ simulation results of different scenarios,ecological land(LCZ A/B/D/G)was effectively protected in the SSP126 scenario,and the expansion of building land(LCZ 1-6、LCZ 8-10)in the main urban area was effectively controlled.Second,in the SSP245 scenario,economic development is also taken into account with ecological protection,and the area of ecological land(LCZ A/B/D/G)only slightly decreases despite the obvious expansion of building land.Finally,in the SSP585 scenario,the expansion of building land(LCZ 1-6、LCZ 8-10)and arable land(LCZ D)is very significant,and forest land(LCZ A,LCZ B)and water bodies(LCZ G)in the mid and late 21st century are seriously damaged.
land simulationlocal climate zonesSSP-RCP scenariosMarkov-MOP-FLUS modelspatiotemporal simulation
NETL
NSTL
万方数据
