Objective Urban areas are increasingly vulnerable to compound rainstorm and heatwave (CRH) disaster risks. Existing research primarily treats rainstorms and heatwaves as isolated risks, resulting in a limited understanding of CRH dynamics and insufficient mitigation strategies. While ecosystem-based disaster risk reduction (Eco-DRR) offers adaptive solutions for multiple disasters, its application to CRH remains underdeveloped. Key challenges include methodological gaps in CRH risk assessment and Eco-DRR application in the planning of disaster risk reduction.

Methods This research develops a planning framework grounded in Eco-DRR theory to address CRH disaster risks. First, the research employs a risk assessment methodology driven by multi-source data to overcome the constraints of traditional single-disaster assessment approaches. The research utilizes daily precipitation and maximum temperature data from Shanghai meteorological stations (2011–2023) to identify CRH events using a maximum temporal interval criterion. Subsequently, disaster records of rainstorms and heatwaves within the event time window are extracted as target variables, while raster data of climatic, topographic, geomorphic, and hydrological influencing factors are derived using ENVI and ArcGIS tools as explanatory variables, forming CRH disaster datasets for training a random forest model. The datasets are partitioned into training and testing sets at a 7:3 ratio. The probability of disaster event occurrence is calculated on a grid-by-grid basis. Disaster risks are classified into high, medium, and low levels using the natural breaks classification method (Jenks), visualized for CRH risks on the ArcGIS platform, and ultimately integrated into a bivariate spatial distribution map through a compound risk matrix. Second, Eco-DRR principles are systematically integrated into territorial spatial planning systems to transition from reactive single-disaster mitigation to proactive resilience-driven strategies. The systematic integration of Eco-DRR theory into the aforesaid planning framework establishes an implementation logic of “risk assessment – planning objectives – support system – spatial configuration – management measures” across five core components. Based on the above, the research proposes the following specific pathways. 1) Resilience goal setting: Defining township/subdistrict-level risk zoning and Eco-DRR targets based on citywide compound risk assessment results. 2) Support system development: Constructing an Eco-DRR support system incorporating mitigation and adaptation strategies. 3) Spatial configuration optimization: Determining spatial allocation schemes for Eco-DRR support elements guided by risk assessment outcomes. 4) Hierarchical management implementation: Coordinating management needs for transition between routine and emergency states under the “risk types – spatial features – planning objectives – management hierarchy” framework. Third, horizontal coordination between ecological spaces and comprehensive disaster prevention systems mitigates fragmentation in existing planning frameworks, establishing a replicable model for multi-disaster, multi-system planning for disaster risk reduction. The Eco-DRR theoretical framework resolves conflicts between multiple planning systems by enabling horizontal coordination between ecological spaces and comprehensive disaster prevention planning. Specifically, Eco-DRR is deconstructed into “+ ecology” and “+ disaster prevention” strategies, with “+ ecology” integrated into comprehensive disaster prevention planning, while “+ disaster prevention” is embedded within ecological spatial planning. Eco-DRR’s mitigation and adaptation strategies are implemented, with coordinated ecological and disaster prevention plans serving as the basis for detailed planning.

Results The research adopts random forest models for analysis to identify CRH events and map their spatial distribution in Shanghai. Results show that CRH disasters predominantly occur between May and September, peaking during the plum rain season and summer months. Annual cumulative durations have increased, exceeding 70 days in the past three years. The high-risk zones for compound risks are concentrated in the central urban areas of Hongqiao, Minhang, and Chuansha districts in Shanghai, as well as surrounding new towns, exhibiting spatial characteristics of “central concentration, peripheral dispersion, and local aggregation”. The spatial distribution patterns of compound risks align with urban development trajectories, with pronounced “rain island” and “heat island” effects. Getis-Ord Gi* analysis reveals that risk hotspots (p<0.05) radiate outward from the urban core to surrounding suburban coldspots. Guided by Eco-DRR theory, dual planning interventions are operationalized: 1) “+ disaster prevention” ecological spatial planning optimization: Eco-DRR constraint indicators embodying the “+ disaster prevention” concept are integrated into Shanghai’s ecological spatial support system. High-risk compound CRH zones are identified as Eco-DRR nodes within the green network, restructuring the outer green belt and suburban green ring. Resilience-compatible zoning is applied based on risk levels. 2) “+ ecology” comprehensive disaster prevention planning optimization: Eco-DRR principles guide “+ ecology” disaster mitigation strategies, including restructuring disaster spaces (shelters, evacuation routes, and zoning) and optimizing safety patterns through risk zoning, route upgrades, and facility improvements. CRH risk zoning informs differentiated construction guidelines, with dual-purpose zoning for normal & emergency states.

Conclusion This research aligns with territorial spatial planning mandates to address CRH risks through Eco-DRR mitigation and adaptation strategies, establishing an integrated territorial spatial planning framework for disaster risk reduction. A random forest-based CRH risk assessment model is developed; empirical analysis is conducted in Shanghai to explore planning pathways under the Eco-DRR theory. District-specific resilience objectives are formulated for subdistricts and structured into “+ ecology” and “+ disaster prevention” strategies. This approach fosters horizontal coordination between ecological spaces and disaster mitigation systems, advancing Eco-DRR integration into territorial spatial planning for disaster risk reduction. The planning methodology provides a replicable framework for CRH mitigation and adaptation in eastern coastal cities. Future research should expand applications to diverse compound climate extremes, incorporate advanced modeling techniques for prediction, and deepen investigations into CRH dynamics and blue – green infrastructure effects.