Development and Evaluation of Phenological Change Indices Derived from Time Series of Image Data

Keith R Mccloy

doi:10.3390/rs2112442

Abstract

Changes in vegetation phenology are recognized as being possibly the most important early indicator of the impact of climate change on ecosystems. Time series of image data are accepted as being the best way to map these changes, if we can derive suitable indices from the huge volumes of time series data that make the interpretation of these time series difficult. The problem is thus to derive a set of indices from the time series of image data that depict changes in vegetation phenology in a way that is easy to analyze and use. This problem has been solved by deriving a set of six Phenological Change Indices that measure the five ways that the phenological curve of vegetation can change over time. These indices were then tested using simulated data based on sample phenological profiles for a set of land covers and showed that four of the indices measured four of the ways that the phenological profile can change, and two of the indices gave similar results in measuring the last way that the phenological profiles can change. A time series of image data was then used to map the Phenological Change Indices for Eurasia and a sample of land covers was used to relate the changes in phenology to location for each land cover. This work showed that the detected changes in phenology are similar to those found in other papers. The benefit of these indices is that we can now analyze changes in phenology in a much more detailed and accurate way than has been possible until now.

Highlights

Phenology is defined as “the study of the timing of recurring biological events, the causes of their timing with regard to biotic and abiotic forces, and the interrelationship among phases of the same or different species” [1]
A time series of image data was used to map the Phenological Change Indices for Eurasia and a sample of land covers was used to relate the changes in phenology to location for each land cover
A simulation study was conducted to verify how varying three factors will affect the phenologies of selected covers and the corresponding changes in the Phenological Change Indices

Summary

Background

Phenology is defined as “the study of the timing of recurring biological events, the causes of their timing with regard to biotic and abiotic forces, and the interrelationship among phases of the same or different species” [1]. For example Ziziphus mauritiana and Acacia senegal are categorized as being deciduous species which have a vegetative phase that coincides with that of the herbaceous layer, whilst other species such as Balanites aegyptiaca are evergreen with new growth phenophases occurring earlier than for the herbaceous layer and the level of green leaf only becoming low due to the presence of predators [10] These processes are likely to explain the variation in temporal responses of different species noted by [5] who have reported that the existing calendar guilds of plants (or associations of plants that exhibit similar phenophase timing) are changing, causing changes in the inter-specific relationships between plants in the Iberian ecosystems. Phenological Change Indices (PCIs) are defined and described and some of the changes detected in the PCIs across Eurasia are discussed

Data Sets Used in This Analysis

Derivation of the Indices

Testing the Derived Phenological Change Indices Using Simulating Data

Results of this Analysis

Discussion