SIE-Climate: A methodological and technological tool for predicting local climate variability in managing socio-ecological systems

Climate variability, as an element of uncertainty in water management, affects community, sectoral, and individual decision-making. Long-range prediction models are tools that offer the potential for integration and joint analysis with the hydrological, hydrodynamic, and management response of the socio-ecological systems to which they are linked. The main objective of this article is to present a seasonal climate prediction model, the open-source algorithm SIE-Climate, whose application consists of three phases (exploration, development, and evaluation), and to describe its application to the Lake Sochagota socio-ecological system (Paipa, Boyaca, Colombia). The K-nearest neighbours method is used when defining a target matrix that represents and integrates macro- and micro-climatic phenomena (Oceanic Nino Index, local temperature, and local rainfall) to identify periods of similar climatic behaviour. Considering a 1-year horizon and management purposes the tool is calibrated and validated in periods with and without climatic anomalies (2000-2018), giving reliable adjustment results (RSME:4.86; R-2: 0.95; PBIAS: -8.89%; EFF: 0.85). SIE-Climate can be adapted to various contexts, variables of interest, and temporal and spatial scales, with an appropriate technological and computational cost for regional water management.