Partitioning savanna tree and grass leaf area index (LAI) from MODIS aggregates in Sub-Saharan Africa

Milkah Njoki Kahiu, South Dakota State University, milkah.kahiu@sdstate.edu
Niall Hanan, South Dakota State University, niall.hanan@sdstate.edu (Presenter)

Leaf area index (LAI) generated from MODIS Terra and Aqua satellites has been widely used for monitoring vegetation dynamics, analysis of phenological patterns, and modeling ecosystem carbon, water and biogeochemical dynamics. For mixed tree-grass systems, however, MODIS LAI provides no information on the relative importance of the woody and herbaceous components in determining the aggregate LAI estimate. In light of this we have embarked on a woody and herbaceous LAI separation exercise using MODIS LAI for sub-Saharan Africa, a region that includes some of the largest mixed tree-grass savannas and dry deciduous woodland systems in the world. Our approach uses new data on slowly varying canopy structure (tree cover) to provide a key constraint in the separation of seasonally varying woody and herbaceous LAI.

Our objectives are to (i) showcase the new methodology for LAI partitioning being developed in this project and (ii) use an analysis of wild-fire patterns in Africa to demonstrate the utility of partitioned herbaceous and woody leaf area index for more detailed study of vegetation dynamics, ecosystem processes and provision of ecosystem goods and services at regional and continental scales.

Previous analysis of satellite-based fire burned area and active burning in tropical savannas reveal a close correlation with satellite-based estimates of total NPP in drier savannas, and apparent limitation by rainfall (fuel moisture) in wetter systems. However, these analyses of fire frequency and extent at continental scales ignore the different roles played by the herbaceous and woody vegetation components in promoting and/or suppressing fire ignition and spread. We hypothesize that, since herbaceous vegetation provides the primary fuel, fire frequency and burn area should correlate more closely with estimates of maximum herbaceous LAI and leaf area duration, than it does with aggregate LAI. Similarly, while fire patterns may correlate with patterns of total LAI across Africa, the relationship will be confounded by variations in tree cover. Using partitioned herbaceous LAI as a proxy for herbaceous fuel load, we show how Africa’s savanna fires correlate better with herbaceous leaf area than with aggregate (total) LAI.

Presentation Type:  Poster

Session:  General Contributions   (Tue 4:35 PM)

Associated Project(s): 

• Hanan, Niall: Partitioning Savanna Tree and Grass LAI and fPAR from MODIS and VIIRS Aggregates: Methods, Validation and Applications ...details
• Hanan, Niall: Partitioning savanna tree and grass LAI and fPAR from MODIS and VIIRS aggregates: methods, validation and applications. ...details

Poster Location ID: 160