Deriving and validating Leaf Area Index (LAI) at multiple spatial scales through lidar remote sensing: A case study in Sierra National Forest, CA

Hao Tang, Matthew Brolly, Feng Zhao, Alan Strahler, Crystal Schaaf, Sangram Ganguly, Gong Zhang, Ralph Dubayah

Research output: Contribution to journalArticlepeer-review


Increasing the accuracy and spatial coverage of Leaf Area Index (LAI) values is an important part of any attempt to successfully model global atmosphere/biosphere interactions. It is further a fundamental parameter in land surface processes and Earth system climate models. Remote sensing methods offer an opportunity to improve on each of these requirements but are typically limited by the necessity for validation using labor intensive and sparsely collected in situ measurements. In this paper we present the results of an intercomparative study of ground-based, airborne and spaceborne retrievals of total LAI over the conifer-dominated forests of Sierra Nevada in California. The efficacy of LVIS (Laser Vegetation Imaging Sensor) airborne waveform lidar LAI measurements (total and vertical profile) has previously been validated at the site specific level using destructive sampling. We also explore the efficacy of ground based measurements obtained from hemispherical photography, LAI-2000, and ground based lidar, acknowledging discrepancies existing between the systems and collected data. We highlight their use and role in validating the relationship between ground and airborne estimates of total LAI (LVIS LAI correlation with i) hemispherical photographs,r2=0.80, ii) LAI-2000,r2=0.85, and iii) terrestrial lidar,r2=0.76. The existence of such relationships offers immediate implications for LAI estimation where LVIS data is available, creating the potential to obtain, not only total LAI values but also corresponding vertical LAI distributions from a ground validated source previously unobtainable at this spatial scale. The ability to validate airborne lidar LAI data collected at different spatial scales to the available ground measurements allows further upscaled validation using global lidar datasets provided by spaceborne lidar, such as the Geoscience Laser Altimeter System (GLAS). In the absence of adequate ground validation plots coincident with GLAS footprints, GLAS LAI validation is examined using geographically limited but spatially continuous LVIS data. Under favorable conditions, significantly the absence of slopes greater than ~20°, the comparison between LVIS and GLAS LAI values obtained using a recursive algorithm constrained by independently validated LAI limits exposes the capability of GLAS as an accurate standalone LAI sensor (r2=0.69, bias=−0.05 and RMSE=0.33). The correlation comparison between LVIS and GLAS LAI estimates not only significantly exceed those associated with equivalent space borne passive remote sensing datasets, such as MODIS (r2=0.20, bias=−0.16 and RMSE=0.67) but also offers significant advantages to future research including the prospective validation of regional and global LAI products and data comparison with ecosystem model inputs. The encountered effectiveness of these relationships allows the implementation of a scaling-up strategy where ground-based LAI observations are related to aircraft observations of LAI, which in turn are used to validate GLAS LAI derived from coincident data. Successful implementation of this strategy paves the way for the future recovery of vertical LAI profiles on a global scale and opens up the potential for fusion studies to incorporate widely available and spatially abundant passive optical datasets.
Original languageEnglish
Pages (from-to)131-141
Number of pages11
JournalRemote Sensing of Environment
Publication statusPublished - 21 Jan 2014


  • Lidar
  • LAI
  • Sierra National Forest
  • LVIS
  • GLAS
  • Echidna


Dive into the research topics of 'Deriving and validating Leaf Area Index (LAI) at multiple spatial scales through lidar remote sensing: A case study in Sierra National Forest, CA'. Together they form a unique fingerprint.

Cite this