Trait–Climate Relationships Are Comparable for Modern Pollen Assemblages and Extant Vegetation Plots

Aim Reliable climate simulations require comparison with robust reconstructions of past climates, often based on fossil pollen assemblages. This is particularly critical in drylands, where climate models frequently diverge. Yet, traditional pollen‐based reconstructions often overlook the mechanistic links between climate and plant functioning, complicating their interpretation. Integrating functional traits into pollen analyses may help identify and reduce biases in climate reconstructions. Location Arid Central Asia. Time Period Present and Holocene. Major Taxa Studied Spermatophytes (pollen‐producing plants). Methods Here, we combine plant functional traits, modern surface pollen assemblages and vegetation plots to test whether assemblage‐level trait–climate relationships are consistent between pollen and vegetation data. Two levels of taxonomic resolution of pollen identification ( fine vs. coarse ) were tested. The biogeographical distributions of plant traits derived from pollen and vegetation were examined, and community‐weighted means of traits were compared between the pollen and vegetation datasets to determine: (1) whether similar trait–climate relationships can be detected in both and (2) which of these relationships are strongest. Results Leaf area, leaf nitrogen content per unit mass, specific leaf area and stem specific density (SSD) showed consistent spatial patterns across pollen and vegetation datasets. The trait–climate relationships were moderately correlated (ca. r of 0.6) between pollen and vegetation data, and the directions of correlations were similar for 85% of the individual relationships. Both fine and coarse taxonomic resolutions capture broad vegetation functional responses to climate, whereas finer pollen identification improves the accuracy of trait–climate relationships. Mean annual precipitation and cold‐quarter temperature and precipitation were the climatic variables most strongly related to traits. Leaf area correlated positively with the three climate parameters, while plant height and SSD correlated positively with annual precipitation. Main Conclusions Our findings support the use of pollen‐derived leaf area, plant height and SSD to track past vegetation structure and improve the reliability of pollen‐based climate reconstructions. Expanding trait databases to better represent water‐stress tolerance traits may strengthen this approach.