Towards a new definition of growing season length

1
Towards a new definition of growing season length

• Michael White
• Utah State University

2
Overview

• Growing season length definitions
• Motivations for current research
• Concept for long-term terrestrial diagnostic
  monitoring

3
Growing season length definitions

• Frost-free duration
• Last spring to first fall frost
• Calendar definition
• Days with average temperature gt 0 or 5C
• Budburst to leaf senescence (or related physical
  metrics)
• Carbon uptake period
• Derived from remote sensing metrics

4
Problems

• Definitions are often mutually exclusive or show
  different ordinal relationships
• Yet the same term is used

5
FPAR greater than 0.5
frost-free days
carbon uptake period
canopy duration
From White, M. A., N. Brunsell, and M. D.
Schwartz, Vegetation phenology in global change
studies, in Phenology An Integrative
Environmental Science, edited by M. D. Schwartz,
pp. 453-466, Kluwer Academic Publishers, New
York, NY, 2003.
6
From White, M. A., S. W. Running, and P. E.
Thornton, The impact of growing-season length
variability on carbon assimilation and
evapotranspiration over 88 years in the eastern
U.S. deciduous forest, Int. J. Biometeorol., 42,
139-145, 1999.
7
From White, M. A., and R. R. Nemani, Canopy
duration has little effect on annual carbon
storage in the eastern United States broad leaf
forest, Global Change Biol., 9, 976-972, 2003.
8
From White, M. A., and R. R. Nemani, Soil water
forecasting in the continental United States
relative forcing by meteorology versus leaf area
index and the effects of meteorological forecast
errors, Can. J. Remote Sens., 30, 717-730, 2004.
9
Uses for growing season length

• Basic biology and physiology
• Prognostic modeling in carbon cycle and climate
  models
• Short term forecasting for agriculture, fire,
  irrigation, famine, etc.
• Education
• Understanding the influences of climate cycles
  (ENSO, PDO, NAO)
• Long-term terrestrial diagnostic monitoring

10
Long-term terrestrial biospheric monitoring

• No specific events or dates
• No annual information
• Should be multivariate
• any one metric likely to be incomplete or
  inappropriate

11
Implement ecological niche concepts

• Combine two commonly used growing season length
  metrics
• Temperature
• Remote sensing measure of greenness

12
Temperature data

• NCEP-DOE Reanalysis II 2m Air Temperature
• T62 resolution (2 degrees)
• Reprojected from Gaussian to 0.25 with NCAR
  Compute Language

13
Fundamental GSL

• Days with average temperature gt 5C

14
(No Transcript)
15
Remote sensing

• Boston University MODIS MOD15 FPAR global 8-day
  composites
• 2001 - 2005
• Sinusoidal, reprojected to 0.25 geographic

16
FPAR GSL

• Find 2001-2005 midpoint between annual low and
  high FPAR values
• Days with FPAR gt threshold FPAR

17
(No Transcript)
18
(No Transcript)
19
Realized GSL

• The intersection of the fundamental and FPAR GSL
• Days greater than 0.5C and above the FPAR
  threshold
• Both warm and absorbing radiation - also good
  proxy for moisture availability

20
(No Transcript)
21
Why bother?
22
(No Transcript)
23
(No Transcript)
24
Diagnostic monitoring

• If longer fundamental GSL means longer realized
  GSL
• 2001-2005 sorted ranking of fundamental and
  realized GSL should have high correlation

25
(No Transcript)
26
Future plans

• Incorporate LST - a humidity proxy
• Analyze phenoregions - groups of phenologically
  and climatically similar pixels

27
Conclusions

• Long-term terrestrial monitoring should ignore
  interannual variability and specific phenological
  events
• Combine remote sensing and climate -
  simultaneously track fundamental and realized GSL