Quantifying Biome specific relationships and Monsoon Event responses

1
Quantifying Biome specific relationships and
Monsoon Event responses
Using LAI , NDVI, PRI, and net ecosystem exchange
in desert grassland and shrubland on the Sevilleta

• Maxine Paul
• Marcy Litvak, Mentor

Photo credit Laura Sanchez Jardon
2
The Big Picture Remote Sensing of Ecosystems

• Satellites measure NDVI (Normalized Difference
  Vegetation index) and LAI (Leaf Area Index)
• Measurements relied upon in ecosystem modeling
• We find it questionable that they are looking at
  factors about an ecosystem on such a large scale
  (1 pixel1km!)

http//svs.gsfc.nasa.gov/stories/drought/images/na
_ndvi0086_sm.jpg
Ecosystems such as the grassland and shrubland on
the sevilleta show lots of heterogeneity on the
ground, are the satellite measurements averages
missing something?
3
The Medium Picture

• Marcy Litvaks lab uses flux towers set up at 2
  locations on the Sevilleta measuring Net
  Ecosystem Exchange of Carbon (NEE),
  meteorological variables, and optical vegetation
  indexes of NDVI, and PRI over a scale of lt 0.5km2

4
The Little Picture

• This summer I have measured NDVI, PRI
  (Photochemical Radiation Index) and LAI on small
  scales (lt1m2 plots) measuring 60cm above ground
• I used a portable sensor developed by Lee
  Vierling at the University of Idaho with a
  Campbell Scientific datalogger (CR23X)

5
Research Questions

• Is there a relationship on the ground between
  NDVI , LAI, and PRI in grassland and shrubland
  biomes? How do they correspond with
  precipitation?
• How well do optical sensing methods compare to
  traditional biomass harvesting methods for LAI
  measurement?
• How well do plot scale measurements compare to
  full ecosystem carbon uptake? How does this
  correspond with precipitation?

6
Grassland and Shrubland ecosystems on the
Sevilleta

• Plant species that dominate the ecosystems
• Creosote in shrubland Larrea tridentata
• Blue Grama (Bouteloua gracilis),and Black Grama
  (Bouteloua eriopoda) in grassland, which use C4
  photosynthesis and are perennial.

• Larrea tridentata, C3 photosynthesis

7
Locations of CO2 flux towers and my transects
Light green grassland site Dark
green shrubland site
8
Transects

• Set up in each biome
• Three 100m transects, one 50 m transect for
  biomass harvesting/allometry.
• Little squares represent a plot every 10 m.
• Laid out in cover of ecosystem tower
• At each plot I measured NDVI and PRI for 5
  minutes
• Quantified covers with quadrat for each
  (understanding of the make-up of each plot, soil,
  grass, shrub, rock)

100 m
N
50 m
9
LAI Methods

• Measurements below and above canopy at each plot
  (lt1m2) along transects at Grassland and Shrubland
  CO2 Flux tower locations
• using a Decagon PAR/LAI ceptometer
• Ratio of how much photosynthetically active
  radiation (PAR) gets through the canopy

Taking LAI measurements with ceptometer
Photo credit Jennifer Johnson
10
NDVI/PRI methods

• The specific wavelengths correspond to pigments
  of plant tissue responsible for photosynthesis
• PRI
• NDVI, corresponds to absorption of
  chlorophyll a (also a band at 800nm, not on
  graph)
• PRI corresponds to photosynthetic light use
  efficiency, equivalent to the rate CO2 uptake by
  foliage per unit energy absorbed and carotenoid
  pigments

NDVI (p800-p675)/ (p800p675) PRI
(p531-p570)/(p531p570) P800 Near Infrared
Radiation (NIR) spectrum P675Red spectrum
11
Net Ecosystem Exchange of Carbon

• NEE Net Ecosystem Exchange of CO2 direct
  measurement of exchange of CO2 between atmosphere
  and ecosystem
• NEEGPP (RplantRhetertrophic). When NEE is
  positive, more carbon is released into atmosphere
  (carbon source). When NEE is negative, more
  carbon is sequestred by ecosystem (carbon sink)

Eddy covariance instruments on towers measure the
integrated fluxes of all of the biota in the
towers range.
Flux tower instruments over a forest
12
1) Is there a relationship on the ground between
NDVI and LAI in these two biomes?

• These trends are describing both biomes together
• Barely negative, very weak correlation in mid
  june. NDVI and LAI not strongly related to one
  another.
• Positive slightly stronger correlation in mid
  July. As LAI increases, NDVI increases.
• Relationship becomes apparent after monsoon
  season has started (july)
• As greenness increases, relationship strength
  increases.

13
Shrubland v Grassland

• Top graph shows shrubland relationship
• Between LAI and NDVI
• Points more spread out than grass,
• Positive trend apparent
• Bottom graph shows grassland relationship
• Between LAI and NDVI
• Points more clumped
• Positive trend apparent
• Shrubland shows a slightly stronger
• relationship over the whole summer
• Between LAI and NDVI.
• Possibly due to wider range of LAI (shrubs
  taller and larger than grasses)
• Points with low NDVI but LAI of around .8 are
  result of patchy areas where a plot has one
  shrub/group of grass but mostly bare ground

14
2) How well do sensing methods compare to
traditional biomass harvesting methods for LAI
measurement?

• Shrubland Biomass (allometry) v LAI (ceptometer)
• Strong positive relationship
• Possible saturation or leveling out
• Not sufficient data points to draw strong
  conclusions

15
Grassland LAI and Biomass
We didnt find correlation between harvested
biomass in the grassland and LAI measured by the
ceptometer. (on left, r2 value of .01) We also
compared live biomass harvested in the grassland
to NDVI, also very little correlation (r2.05).
16
3) How well do plot scale measurements compare to
full ecosystem Carbon uptake? How does this
correspond with precipitation?
Shrubland
Grassland
17
Grassland responses
Shrubland responses
NEE of C and LAI
NEE of C and NDVI
NEE of C and PRI
18
Conclusions

• Is there a relationship on the ground between
  NDVI and LAI in these two biomes?
• In both biomes, after the monsoon season has
  started, there is a positive correlation between
  NDVI and LAI. As biomass increases, Normalized
  difference vegetation index becomes more
  positive.
• In the shrubland the relationship between NDVI
  and LAI is slightly stronger. One possibility is
  that it is due to larger range of LAI for
  creosote than for gramas.
• My results suggest there are stronger
  relationships in mid-summer in both biomes,
  compared to early summer. This correlates with an
  increase in rainfall and available water for
  photosynthesis.

19
Conclusions

• 2) How well do optical sensing methods compare to
  traditional biomass harvesting methods for LAI
  measurement?
• For shrubland, allometric biomass was positively
  related to LAI, and had strong correlation. In
  this study ceptometer use is similar to allometry
  of Creosote.
• For grassland, no strong correlation between
  biomass harvesting and LAI by ceptometer.
  Results suggest that optical measurements are
  better for understanding relationship between LAI
  and NDVI in grassland.

20
Conclusions

• 3) How well do plot scale measurements compare to
  full ecosystems Carbon uptake? How does this
  correspond with precipitation?
• Because grass seems to respond quicker to
  rainfall events, the grassland LAI, NDVI show a
  stronger correlation to carbon uptake
• In the shrubland, there is an increase in LAI
  most likely in response to precipitation, however
  the relationship to C flux is less clear. With
  partitioning out GPP and respiration, we expect a
  stronger relationship.
• PRI, NDVI, LAI in both biomes show a seasonal
  progression corresponding to precipitation

21
Future Work, Connections

• Southwest US undergoing climate change
  semi-arid ecosystems are particularly sensitive
  to changes in climate
• Over the year I will continue to take NDVI, LAI,
  and analyze PRI measurements and compare the
  little to medium to big pictures
• -NEE chamber added into thesis, more PRI analysis
• -Look for me at the Earth and Environmental
  Science Senior Thesis presentation at Columbia
  University in the spring in NYC

22
References

• -L. Fan, Y. Gao, H. Bruck, Ch. Bernhofer,
  Investigating the relationship between NDVI and
  LAI in semi-arid grassland in inner Mongolia
  using in-situ measurements. Theoretical and
  Applied Climatology. March, 2007
• -M. Van Wijk, M. Williams, Opitical Instruments
  for Measuring Leaf Area Index in Low Vegetation
  Application in Artic Ecosystems, Ecological
  Applications, 2005.
• - National Aeronautic and Space Administration,
  Goddard Space Flight Center, One Year Vegetation
  Average- NDVI measures plant growth
  http//svs.gsfc.nasa.gov/stories/drought/

23
Acknowledgements

• -Marcy Litvak
• -Andrew Fox
• -Litvak Lab
• -Jennifer Johnson
• -Cesar Coronado, Frankie Reyes, Giomara LaQuay,
  Diana Guzman, Laura Sanchez Jardon, Ernest
  Herrera, all interns and REUs
• -Sevilleta LTER
• -National Science Foundation
• -UNM Biology

24
Questions?