Phenological environmental assessment indicators – proposing an international standard

Phenological change is the most obvious ecological effect of climate change on ecosystem properties, processes and services. Estimates of phenological change are increasingly being used in environmental, ecological and climate change assessments, both at the international (e.g. IPCC WG2) and national levels (e.g. NCA, USA). Despite of this, there is no international standard for the analytical procedures (A), no standard for the pheno-metrics (PM), no standard for baselines (B), and no standard for data quality classification (QC). The phenology monitoring community represents a wide range of systems – from professional to volunteer observers, from developing systems to those that have been working for over a century, from spatially large networks to point observations, from systems with a focus on agriculture to those more oriented to wildlife. To be able to merge data from these different systems and to provide widely applicable pheno-metric products, there is a great need for a common standard. We present a proposal for a phenological environmental assessment indicator standard. Standardized quality classification (QC) includes four levels: (QC1) Observer self-validation. (QC2) Error check (e.g. of required phase order) leading to flagging of erroneous data to exclude from the analysis. (QC3) Outlier check of both single observations and inter-annual variation leading to parallel analyses with/without flagged outliers. Outlier check should be based on (i) models using historical data and (ii) models using present-year neighboring observations. (QC4) Image-based confirmation that can overrun (QC3) flagging decisions. As a standardized baseline (B), we propose estimated phenology data for the climatological standard periods (e.g. 1961-90), using a phenology model that has been calibrated regionally. We propose two different pheno-metrics: The first (PM1) is simply number of days deviation from the baseline, B, and the second (PM2) is the proportion of deviation compared to the local, meteorologically defined, vegetation period. Analysis could be (A1n) single-year estimates, (A2n) short period averages based on at least 5 years of data, or (A3n) long-term averages based on at least 30 years of consecutive data. A1n-A3n could be presented as (Am1) point estimates or (Am2) spatial averages based on at least 10 separate stations with data overlapping in time. Indices, combining different species+phase combinations, can be developed if all included combinations can be referred to the same, minimum level of (Amn). We present an example of pheno-metrics (PM) developed for the Swedish Environmental Assessment and Environmental Objectives, and how we applied the proposed quality classification (QC), two different types of baseline (B), and analytical procedures (A) using historical and present-day data from the Swedish National Phenology Network.