Global greening phase patterns

1) Sensor results

2) Model results

Assessing the seasonality of multi-source fAPAR time-seriesCarola Dahlke, Alexander Loew, and Christian Reick

Max Planck Institute for Meteorology, KlimaCampus, Hamburgcontact: carola.weiss@zmaw.de

A new method to compare global vegetation greening phase dynamics

The fraction of Absorbed Photosynthetically Active Radiation (fAPAR) is an essential diagnostic variable to investigate the temporal and spatial dynamics of the terrestrial biosphere. We introduce a new algorithm that allows for the robust identification of seasonal signals from multi-annual time series with a special focus on the difference in the seasonal phases of the characteristic signal.

Robust greening phase analysis

Data

Pattern comparison

1) Sensor results Combined analysis results

2) Model results

Figure 2: fAPAR values from SeaWiFS time series (examples from 1998)

fAP

AR

4km2003 - 2009MODIS combined

Boston University [4]

0.5 degree

Spatial

resolution

1998 – 2005

1993 – 2000

Time series Temporal

resolution

Sensor

SeaWiFSJoint Research Center [3]

monthly

AVHRR Pathfinder

Boston University [2]

daily1 degree1960 – 2009Global Analysis

MeteoSwiss Zürich [6]

monthly4 degree

2 degree

Spatial

resolution

1980 – 2009

1985 – 2006

Time series Temporal

resolution

Model

ORCHIDEEMPI-BGC, Jena [7]

bimonthlyJSBACH

MPI-M, Hamburg [5]

By vector normalisation and integral calculation [1], we isolate the seasonal signal from any amplitude aspects that are influenced by the inconsistencies of the various data sets, derived from fAPAR time series from remote sensing sensors, and from climate models.

Figure 1: Phasing calculation of a sample normalised annual fAPAR curve

References[1] Weiss, C. et al. (2011): Robust identification of greening phase patterns from remote sensing vegetation products. Submitted to Journal of Climate. [2] Myneni, R. et al. (1997): Estimation of global leaf area index and absorbed par using radiative transfer models. IEEE Transactions on Geoscience and Remote Sensing 35, 1380-1393[3] Gobron, N. et al. (2002): Sea Wide Field-of-View Sensor (SeaWiFS): An optimized FAPAR Algorithm.Technical report. Theoretical Basis Document by Institute for Environment and Sustainability Joint Research Centre, I21020 Ispra (VA), Italy[4] Knyazikhin, Y. et al. (1999): MODIS Leaf Area Index (LAI) and Fraction of Photosynthetically Active Radiation Absorbed by Vegetation (FPAR) Product (MOD15). Technical Report. Theoretical Basis Document by Institute for Environment and Sustainability JRC.[5] Raddatz T.J. et al. (2007): Will the tropical land biosphere dominate the climate-carbon cycle feedback during the twenty-first century? Climate Dynamics 29, 565-574.[6] Stoeckli, R. et al. (2011): A Global Reanalysis of Vegetation Phenology. Journal of Geophysical Research, 116.[7] Krinner, G. et al. (2005): A dynamic global vegetation model for studies of the coupled atmosphere-biosphere system, Global Biogeochem. Cycles, 19, GB1015.

6 5 4 3 2 1 0 MODIS vs. SeaWiFS

AVHRR vs. SeaWiFS

0

1 2

3

4

5 60 1 2 3 4 5 6

AVHRR vs. MODIS

months

mon

ths

months

The pattern comparison results show consistent global spatio-temporal patterns significant at the 95% confidence level. Based on the monthly resolved data sets that have been evaluated in this study, no remarkable shifts are visible. Shifts stay in the range of +/- 1 month, which is the expected minimum shift for monthly resolved data.

AcknowledgementsRemote sensing data was provided by Boston University and the European Joint Research Center (JRC), model data was partly provided by Sönke Zähle from the MPI-BCG in Jena, and by Reto Stoeckli from the ETH Zürich, which is gratefully acknowledged. This study was partly supported through the Cluster of Excellence 'CliSAP' (EXC177), University of Hamburg, funded through the German Science Foundation (DFG).

The robust greening phase pattern algorithm is ideal for the assessment of seasonal processes simulated by the vegetation components of climate models.Regions and land cover types, where the seasonality of the models agree with the remote sensing data sets can be detected as well as regions, where models and observations disagree.

Figure 3: (left) Fraction (mode value)-per-latitude diagrams per sensor time series; (right) sample greening phase patterns for 30% fraction of total greening phase

Figure 6: Combination of the phase comparison of three sensors: White regions display no shift between the sensors, i.e. very good agreement, coloured regions show high agreement between two sensors and low agreement with the third sensor.

Figure 4: (left) Fraction (mode value)-per-latitude diagrams per model time series; (right) sample greening phase patterns for 30% fraction of total greening phase

Figure 5: (left) Shifts in time between the greening phase results from two sensors: fraction (mode values)- per-latitude diagrams;(right) sample shift maps show the comparison of the greening phase patterns for 30% fraction of total greening phase from two sensors

Figure 7: (left) Shifts in time between the greening phase results from a model to a sensor: fraction (mode values)- per-latitude diagrams; (right) sample shift map showing the comparison of the greening phase patterns for 30% fraction of total greening phase from a model to a sensor;(bottom) Major biome distribution: models mostly perfom with high agreement in Tundra, Taiga and temp. Grassland biomes; seasonality of Savanna and Rainforest biomes needs adjustment.

Greening Phase Analysis

� Robust and fast comparison of seasonality� Applicable to multi-source time

series from various vegetation indices� Allows for the validation of climate models