Spatial convergent cross mapping to detect causal relationships from short time series

Area of Science:

• Ecology
• Complex Systems Analysis
• Time Series Analysis

Background:

• Ecological studies often have short time series (fewer than ten observations).
• Spatial replication (plot replication) is common in ecological data.
• Existing causal detection methods require longer time series (around 30 observations).

Purpose of the Study:

• To develop a novel method for detecting causal relationships in ecological systems with short, spatially replicated time series.
• To combine convergent cross mapping (CCM) and dewdrop regression to leverage spatial replication.
• To introduce multispatial CCM for causality assessment in data-limited ecological scenarios.

Main Methods:

• Developed multispatial convergent cross mapping (CCM) by integrating CCM and dewdrop regression.
• Utilized spatial replication to enhance causal inference from short time series.
• Tested the method using simulated and real-world ecological datasets.

Main Results:

• Multispatial CCM successfully identified causal relationships with as few as five sequential observations.
• The technique demonstrated robustness against process noise and observation error.
• Effectiveness was validated on both simulated and empirical ecological data.

Conclusions:

• Multispatial CCM provides a powerful tool for establishing causality in ecological systems lacking long time series.
• This method is particularly valuable for systems where experimental manipulation is difficult.
• The multispatialCCM R package is now available for researchers to apply this technique.