Dear, Huimin Lei, In such situation i suggest you to use an instrument called CI 110 Digital plant canopy analyzer.

There was a publication last year,

http://www.eurekalert.org/pub_releases/2011-06/asfh-nnm062611.php

http://hortsci.ashspublications.org/content/45/10/1459.abstract

Regards,

Joachim

Hi Huimin Lei, The canopy seems short, so I would try using the Li -COR's LAI-2000 with 2 wands - 1 above the canopy and one below.

I'm not sure the methods that use light extinction to get at effective LAI (e.g., licor 2000, hemi photos) would work in such an environment. Leaf area index is half of the area of leafy material in square meters per square meter of ground. So, to get this you could: (1) Cut up a few example shrubs and grasses and actually lay all leaf material out. Calculate an average/reference leaf area from these destroyed trees. Then (2) do a 500 m x 500 m census of plant cover that involves mapping where each type of plant is. Alternatively you could use aerial photography to map this. The bigger the sampled area, the better your estimate. (3) Extrapolate the LAI you associated w/ each type of plant cover to the locations where that plant exists and do an area-weighted interpolation to get a mean LAI in that sampled 500m x 500m pixel. For example, if the LAI of the shrub is 1 m2/m2 but only takes up 5% of the land area in the sampled area, the 500m pixel LAI will be 0.05. Make sense?

You can try ImageJ (find it on Google, download and install). The use of simple image processing technique is in the time saving and high accuracy. ImageJ in particular will detect both shapes and color values. Its a destructive method, so you need to take leaf samples, put them on white paper, take a picture (with digital camera or mobile phone) and use ImageJ to calculate LAI. Don't forget to specify the dimension of one pixel (the resolution of your camera) in the beginning. I am sure some examples are on the net.

Hope it will help.

To improve estimates within your plots, you could follow the above protocol (destructively sample biomass components,) in order to develop allometric equations that relate leaf area to stem diameter. Something similar could be done for grasses too (area of ground covered - biomass). Remember that SLA / LMA can be used with the ratio of wet : dry leaf mass from biomass data to give fresh leaf area per individual.

With such low vegetation values, why don't you just use veg. fraction (Similar to Kiril's suggestion). You only need a digital camera and it relates to biomass/LAI, especially with low LAI as in your situation. As Evan and Dylan suggests, you can take a few destructive measurements to provide 'ground truth' to your remote estimates. I suggest using a leaf area meter to measure LAI destructively as it is easier than measuring each individual leaf. What are the goals of your study. This would make it easier to suggest different methods because destructive measurements are very time consuming and have a negative impact on the environment.

If you have access to spectroradiometers such as the USB2000 (Ocean Optics) or FieldSpec (ASD) you could use previously published Vegetation Indices to estimate LAI. There are plenty of papers that do this in various vegetation types.

You could use very high resolution multispectral satellite data (i.e., Quickbird, IKONOS) with object-based image analysis to classify types of shrubs. You should then take samples of leaves in each class to estimate LAI. Whit this data, you could adjusts statistical models that relate reflectivity at the object level with LAI. Best regards.

You can try LP-80 Ceptometer which can get directly LAI based on light attenuation in the canopy.It is a non destructive method. Another method see: skyeinstruments.com - a software used to measure leaf area.Try it.

If you plan on using an optical instrument (LAI-2000/2200, LP-80 etc) as some have suggested, be sure to have a large sampling. Very open canopies require more sampling points than a forest.

As expected there are several methods of tree cover analysis. The issues are cost considerations and the acccuracy with which you need the information. I think there is already enough information available.Good luck.

Direct methods require stripping and measuring the foliage of plant canopy samples.Indirect methods of estimating LAI in situ can be divided in two categories: (1) indirect contact LAI measurements such as plumb lines and inclined point quadrats and (2) indirect non-contact measurements.

The method depends on the size of the area you have to investigate. For an area of around 10 x 10 km you could draw homogeneous vegetation polygones with GIS. Then you should measure (using one of the methods above suggested) the average LAI for each vegetation typology and apply to polygones for geometrical analysis. In order to study wider areas you should apply the same method to sample areas and extend results to all surface.

Dear all. Thanks for the suggestions. Besides the methods of measuring leaf area, I think the design of quadrats (plots) is also a critical point. What is the size of a plot? How many plots are needed? How do the plots distribute? I also attach a picture which is from Google earth. The size of the red square in this picture is 1km*1km.

ps: the goal of my study is to provide the true ground-based LAI values for validating the remotely-sensed products.

In the picture, I can see some pattern that it is important to take into account. If you want to deal about the red square, subdivide it in little squares searching for one size that shows the less variance (see "size plot") considering a descriptive variable as "red tones" (for example). Then, with the size selected you got a new population of mini squares, to apply a sampling procedure considering some criteria (precision, cost, etc.). I suggest you a systematic sampling. In the sampling squares apply the methods for LAI estimation.

I'd suggest the use of hemispherical photographs taken at ground level and sampled at a regular basis. This is a relatively quick and non-destructive measure that may fit your purpose. You only need to be sure to take a big enough number of samples to cath all the environmental heterogeneity.

There are quite a few papers describing the methods already mentioned above (hemispherical photographs, LAI2000, Leaf Area Meters, etc.) for sparse vegetation. The widely used normalized difference vegetation index was originally developed in range land, thus probably appropriate in your ecosystem.

Here are a few examples from a google scholar search of "LAI sparse shrubland":

http://www.sciencedirect.com/science/article/pii/S003442570000119X

http://www.sciencedirect.com/science/article/pii/S0034425706000368

http://www.sciencedirect.com/science/article/pii/S0034425702000755

I compared three methods to estimate foliage density (LAI can be easily deduced from this) in the understorey of a tropical rainforest. For details look at:

https://www.researchgate.net/publication/231590135_Evaluation_of_methods_to_estimate_foliage_density_in_the_understorey_of_a_tropical_evergreen_forest

To cut a long story short, errors in estimation using point quadrat or photographic methods were in the same range as those obtained when using methods based on direct beam or diffused light penetration (far more costly equipment required). This of course is not surprising because, these instruments also use methods that assume a poison distribution of leaves.

Article Evaluation of methods to estimate understorey foliage densit...