Leaf area index (LAI) is the ratio of upper leaf surface area to ground area (for broadleaf canopies), or projected conifer needle surface area to ground area (for coniferous plants) for a given unit area. LAI directly quantifies canopy structure, and can be used to predict primary productivity and crop growth. It is commonly used in ecosystem models because it has an important influence on exchanges of energy, water vapor and carbon dioxide between plants and the atmosphere. Hence, many ecosystem process models require LAI as an input variable. LAI can be measured on the ground by harvesting leaf tissue and quantifying the leaf surface area or by various indirect techniques, such as hemispherical photography or the use of optical instruments (Plant Canopy Analyzer, DEMON, ceptometer, etc). However, over large areas it is useful to estimate LAI from remotely sensed images. LAI can be a parameter of interest on its own, or can be used as an input to models of primary productivity and fire dynamics.
LAI can be generated from satellite images using various methods, including:
These methods require correction for atmospheric variation and sometimes require bidirectional reflectance normalization. The images are composited over multiple days (i.e., the value for any given pixel in the final image is taken from the highest-quality readings for that pixel across multiple images) to minimize the impact of atmosphere and screening by clouds or snow. The relationship between satellite measures of reflectance and estimates of LAI will vary depending on the type of vegetation being considered, and thus major land cover type is an important input to calculating LAI. Satellite measurements of reflected radiation are often used to estimate the LAI values that are used as an intermediate variable in models of NPP.
Remote sensing LAI methods generate a map of dimensionless LAI values assigned to each pixel. Values can range from 0 (bare ground) to 6 or more, but since rangeland vegetation is generally sparse, values commonly range from 0-1. A LAI value of 1 means that there is the equivalent of 1 layer of leaves that entirely cover a unit of ground surface area, and less than one means that there is some bare ground between vegetated patches. LAI values over 1 indicate a layered canopy with multiple layers of leaves per unit ground surface area. LAI and fPAR data are commonly packaged together (e.g., MODIS products).
LAI is an important input into many ecosystem models. These models can be used to characterize:
Rangeland applications include:
Remotely sensed LAI estimates are only approximations of true LAI (i.e., LAI measures that would be directly obtained by stripping all leaves from an area and quantifying their surface area per unit ground area). The mathematical models used to calculate LAI vary widely, and each model contains assumptions and requires specific inputs. It is important to understand the model assumptions and assess the suitability of the model based on the available data, how well the model characterizes the vegetation compared to field measurements, and the desired output. Most models work optimally at a particular scale and in a particular ecosystem type, and the application of an existing model to a new location may require changes to the model. LAI is often derived from spectral vegetation indices, such as NDVI, but there is no single equation with a set of coefficients that can be applied to images of different surface types. Estimation of LAI by satellite imaging requires corrections for atmospheric effects, topography and diurnal variations, and values change rapidly throughout the season with changing phenology. LAI estimates from visible/near-infrared images require a cloudless, clear image, and thus LAI values are typically chosen from the best quality images over a multiple day period (often an 8 or 10-day window). For areas that are continually cloudy, the use of radar or lidar may be necessary to assess vegetation characteristics.
Remote sensing of LAI requires an image with a visible band and a near-infrared band.
Calculating LAI requires image processing and statistical/mathematical modeling software.
Global LAI from MODIS imagery (source: Running, S.W. and NTSG. 2002 powerpoint presentation).
Images of the Ouachita Mountains in the southwestern US, showing how different indices can provide different types of information. The top three panels are images from April and the bottom three panels are from May of 2004. The leftmost panels (top and bottom) show a near natural color image, the middle panels show LAI calculated from the images, and the right panels show fPAR (fraction of photosynthetically active radiation). (image source: Short, N. 2009. The Remote Sensing Tutorial, Section 3. Online tutorial)
Relationship between annual NPP and LAI using data from 6 decades and across sites throughout the world (source: Scurlock et al. 2001)
Comparison of remotely sensed LIDAR estimates of LAI to ground measurements of LAI from hemispherical photographs (source: Morsdorf et al. 2006)
You must have an account and be logged in to post or reply to the discussion topics below. Click here to login or register for the site.