Abstract

Satellite-observed leaf area index (LAI) is increasingly being used in climate modeling. In common land surface models, LAI is specified for the vegetated part only. In contrast, satellite LAI is defined for the total area including both vegetated and nonvegetated fractions. Some recent modeling studies and model developments have not noticed this difference, which resulted in improper use of satellite LAI. This paper clarified this issue. A sensitivity test was carried out using a regional model to investigate the impacts of LAI definitions on simulated climates. This study showed that use of satellite LAI without considering the inconsistency in definition caused much smaller LAI values in the model. As a result, partitioning of surface energy into latent and sensible heat fluxes, as well as the model-simulated precipitation, was affected substantially. Overall, improper use of satellite LAI increased the model biases in simulated precipitation.

1. Introduction

Land surface processes impact regional to global climate processes by controlling exchanges of mass, energy, and momentum between ecosystems and the atmosphere (Pielke et al. 2002). Over the past decades, interactions between the biosphere and atmosphere have been widely investigated (e.g., Xue 1997; Betts 2000; Foley et al. 2003; Ge et al. 2007), and the importance of land surface change in global and regional climate change has received increasing attention (e.g., Feddema et al. 2005; Turner et al. 2007). Since Deardorff’s (1978) pioneering work, land surface models (LSMs) have evolved from quite simple treatments of the surface energy, moisture, and momentum exchanges to increasingly complex descriptions (Pitman 2003). Paralleling this development, an increasing number of satellite observations have been used in climate models, either to better parameterize land surface conditions or to more realistically represent land surface changes. Leaf area index (LAI), commonly defined as the one-sided foliage area per unit ground surface area, controls many physical and biological processes such as photosynthesis, transpiration, and rainfall interception. In recent years, many climate modeling studies have used satellite-observed LAI for improving surface characterization or investigating vegetation change impacts due to human activities (Sellers et al. 1994; Chase et al. 1996; Bounoua et al. 2000; Oleson and Bonan 2000; Buermann et al. 2001; Lu and Shuttleworth 2002; van den Hurk et al. 2003; Lawrence and Chase 2007; Kang et al. 2007; Ge et al. 2008). LAI data are primarily from the National Oceanic and Atmospheric Administration’s (NOAA) Advanced Very High Resolution Radiometer (AVHRR) and the National Aeronautics and Space Administration’s (NASA) Moderate Resolution Imaging Spectroradiometer (MODIS) observations. These studies have demonstrated the value of satellite observations for land–climate modeling.

However, there is a definition difference between LAI derived from satellites and LAI specified in LSMs, which has not caught much attention. Improper use of satellite-observed LAI may add extra uncertainties in model simulations. The purpose of this paper is to address this issue and to illustrate its impacts on simulated climate using a regional model.

a. LAI definitions

In LSMs, a given surface grid cell is typically treated as having a specified fraction of fully vegetated surface, defined by vegetation fractional cover (VFC), with the remainder of the surface assumed to be nonvegetated bare soil. This also applies to subgrid patches (or tiles) when the mosaic approach (e.g., Walko et al. 2000) is used for greater surface heterogeneity. LAI is specified only for the vegetated fraction in commonly used LSMs, including the Biosphere–Atmosphere Transfer Scheme (BATS; for LAI definition see Dickinson et al. 1998) and the Simple Biosphere Model (SiB; for LAI definition see Sellers et al. 1986). Many other later developed LSMs also define LAI for the vegetated part only, for example the Land Ecosystem–Atmosphere Feedback model (LEAF; for LAI definition see http://www.atmet.com/html/docs/rams/rams_techman.pdf), the National Center for Atmospheric Research (NCAR) LSM (for LAI definition see Oleson and Bonan 2000). In this paper, LAI for the vegetated part is referred to as “clump” LAI (CLAI).

In contrast, the LAI from satellite observations is ultimately based on reflectance measured over the entire pixel (vegetated and nonvegetated), regardless of methods used (e.g., Sellers et al. 1994; Myneni et al. 1997). Thus, LAI from remote sensing quantifies the amount of foliage area per unit total ground, and this definition remains same when the high-resolution (e.g., 1 km) LAI data are aggregated to resolutions of climate models. In this paper, remote sensing LAI is referred to as RLAI.

To remove the inconsistency, RLAI needs to be divided by VFC to obtain CLAI, which is actually needed in LSMs. It may cause substantial decrease of LAI values if RLAI is directly used for climate modeling, especially when a region has low VFC value, for example, semiarid areas. While some studies have noticed this difference in definition (e.g., Oleson and Bonan 2000; van den Hurk et al. 2003), others used RLAI with less caution. Kang et al. (2007) compared AVHRR-derived RLAI directly to prescribed CLAI in a general circulation model with the Simplified Simple Biosphere Model (SSiB). They found much smaller LAI values compared to survey data. This might have contributed to the reported increase of temperature and decrease of precipitation. The recent development of a regional model, Regional Atmospheric Modeling System (RAMS) version 6.0, includes the feature to use AVHRR-derived RLAI in the LEAF surface model (http://www.atmet.com/html/docs/rams/RT1-leaf2-3.pdf). However, RLAI is adopted without adjustment, which may add undesirable uncertainties to future uses. While there is no evidence to verify the improper use of RLAI, several studies showed much smaller LAI values because of the use of satellite data (e.g., Buermann et al. 2001; Lu and Shuttleworth 2002).

The difference between CLAI and RLAI is illustrated at the regional scale focusing on the East Africa region (Fig. 1). In Fig. 1, a land cover map, Global Land Cover (GLC) 2000 for Africa (Mayaux et al. 2004), is used to illustrate the land cover conditions in this region. This region has a very dynamic landscape: Congo forest west of Lake Victoria, highlands and intensive agricultural activities around Lake Victoria, and semiarid area in the northeast of Kenya. Precipitation pattern is strongly influenced by the migration of the intertropical convergence zone (ITCZ). In general, the period from February to May is the main wet season and the relative dry season for large part of the domain is around June. This study looks at the spatial and temporal dynamics of both CLAI and RLAI for a half year (February–July) in 2003, which is a normal year in terms of total precipitation. Monthly 1-km MODIS LAI data were obtained from the MODIS group at Boston University. VFC data were developed from the 1-km monthly MODIS Enhanced Vegetation Index (EVI; Huete et al. 2002) product, based on the theory of “mosaic pixel” (Ge et al. 2008). In addition, the prescribed LAI in the LEAF land model (version 2) in RAMS is included for comparison.

Fig. 1.

Map of the study area represented by GLC 2000 of 1-km resolution. The large inland water body near the center of the map is Lake Victoria.

Fig. 1.

Map of the study area represented by GLC 2000 of 1-km resolution. The large inland water body near the center of the map is Lake Victoria.

Figure 2 shows the spatial and temporal dynamics of VFC (Fig. 2a), RAMS-prescribed LAI (DEF LAI; Fig. 2b), RLAI from MODIS (Fig. 2c), CLAI (MODIS LAI divided by VFC with a maximum value set to 7; Fig. 2d), and the differences between RLAI and CLAI (Fig. 2e). Only three months of data are shown for the illustration purpose. The resolution is 50 km, which is the grid spacing used for the RAMS simulation (shown in the next section). The 1-km VFC and LAI image data were inserted in RAMS (Ge et al. 2008) and aggregated to patch level in the 50-km grid cell based on land cover types. In Fig. 2, VFC and LAI values shown are for the biggest patches in grid cells.

Fig. 2.

Spatial and temporal comparison of (b) RAMS-prescribed LAI (DEF LAI), (c) MODIS LAI (RLAI), and (d) clump LAI (CLAI). (a) MODIS-derived VFC; (e) the difference between RLAI and CLAI (CLAI − RLAI).

Fig. 2.

Spatial and temporal comparison of (b) RAMS-prescribed LAI (DEF LAI), (c) MODIS LAI (RLAI), and (d) clump LAI (CLAI). (a) MODIS-derived VFC; (e) the difference between RLAI and CLAI (CLAI − RLAI).

In Fig. 2, a substantial difference exists between RLAI and CLAI, and this difference changes with space and time under the influence of the VFC characteristics. In general, CLAI is much higher (greener) than RLAI, especially for those areas with small VFC. As ITCZ moves to the north (e.g., June), the low VFC zone in the eastern part of the domain migrates to the south and affects the spatial pattern of CLAI. DEF LAI does not vary much in space and time. Both RLAI and CLAI are smaller than DEF LAI, but CLAI is closer to DEF LAI as they have consistent definitions. Table 1 lists mean and maximum difference between CLAI and RLAI over the domain (excluding water). Overall, the domain-averaged difference is more than 1 LAI with a maximum difference around 4 LAI. As the region becomes drier (particularly eastern domain), differences between CLAI and RLAI grow substantially.

Table 1.

Mean and maximum LAI differences between CLAI and RLAI (CLAI − RLAI) for six months in 2003. Only land cells were included to calculate these statistics.

Mean and maximum LAI differences between CLAI and RLAI (CLAI − RLAI) for six months in 2003. Only land cells were included to calculate these statistics.
Mean and maximum LAI differences between CLAI and RLAI (CLAI − RLAI) for six months in 2003. Only land cells were included to calculate these statistics.

b. Sensitivity study

A sensitivity test was conducted to investigate the impacts of improper use of LAI on simulated climates. RAMS version 4.4 (Pielke et al. 1992) was used to simulate the period from February to July 2003. January of 2003 was used as the spinup time and thus not considered for analysis. A single grid with 50-km horizontal grid spacing was used for the model domain (Castro et al. 2005). The grid extended over 32 vertical levels with a layer thickness of 80 m near the surface. The model was driven by 6-hourly lateral boundary conditions derived from the National Centers for Environmental Prediction (NCEP) atmospheric reanalysis product (Kalnay et al. 1996). The model time step was 90 s with the output period set to every 3 h. Monthly VFC and LAI data were interpolated linearly to determine daily values. See Ge et al. (2008) for more detailed model configuration.

Two numerical experiments were carried out. In the first experimental run, here called “RLAIRUN,” MODIS VFC and LAI were used directly in RAMS. In the second run, here called “CLAIRUN,” MODIS VFC was used with CLAI for consistency in LAI definition. The GLC 2000 land cover dataset was used in both runs to replace the RAMS default land cover map for better representation of surface conditions in this region (Ge et al. 2008). For this study, only surface fluxes and precipitation were analyzed for illustration purposes. Other variables (e.g., temperature and atmospheric circulation) may need to be examined for a more complete understanding of LAI sensitivity.

Figure 3 shows the surface flux difference between CLAIRUN and RLAIRUN. Dotted contour lines indicate statistically significant changes with a 90% confidence level. Daytime (1200 UTC, or about 2:00 p.m. local time) latent and sensible fluxes were averaged over the six-month time period. CLAIRUN gave much higher latent heat flux in large part of the domain (Fig. 3a). The domain-averaged difference is 3.13 W m−2. This is primarily due to higher LAI values and thus enhanced evapotranspiration. The changes of sensible heat flux exhibit an overall similar pattern to that of latent heat flux except of opposite sign (Fig. 3b). The changes in latent and sensible heat fluxes indicate the important role of LAI in partitioning net radiation to surface heat fluxes. In the areas west and north of Lake Victoria, latent heat flux decreased and sensible heat flux increased in response to higher LAI values. In this case, higher LAI slightly lowered the vegetation albedo, and downward shortwave radiation increased because of reduced cloud cover. These two effects resulted in slight increase in net radiation. Reduced moisture in this area (see Fig. 4 for precipitation change) decreased latent heat flux, and thus sensible heat flux increased accordingly. In the eastern domain, particularly northeast Kenya where moisture is very limited, higher LAI by CLAI did not increase latent heat flux but sensible heat flux decreased significantly. This is attributed to the combined effects of LAI influence on albedo and upward longwave radiation changes due to surface temperature variations (e.g., Oleson and Bonan 2000). In this case, downward shortwave radiation did not change, and higher LAI lowered the surface albedo. But the increased surface (vegetation and soil) temperature increased upward longwave radiation substantially. The net radiation thus decreased. At the same time ground heat increased. This resulted in significant decrease of sensible heat flux.

Fig. 3.

Difference of (a) averaged surface latent heat flux and (b) sensible heat flux between RLAIRUN and CLAIRUN experiments (CLAIRUN − RLAIRUN) (W m−2). Dotted contour lines in (a) and (b) indicate the 90% confidence level from the Student’s t test.

Fig. 3.

Difference of (a) averaged surface latent heat flux and (b) sensible heat flux between RLAIRUN and CLAIRUN experiments (CLAIRUN − RLAIRUN) (W m−2). Dotted contour lines in (a) and (b) indicate the 90% confidence level from the Student’s t test.

Fig. 4.

(top) Difference of accumulated precipitation between CLAIRUN and RLAIRUN and (bottom) a comparison of RAMS simulated (CLAIRUN) and observed (TRMM) accumulated precipitation (mm). Dotted contour lines indicate the 90% confidence level from the Student’s t test of precipitation rates.

Fig. 4.

(top) Difference of accumulated precipitation between CLAIRUN and RLAIRUN and (bottom) a comparison of RAMS simulated (CLAIRUN) and observed (TRMM) accumulated precipitation (mm). Dotted contour lines indicate the 90% confidence level from the Student’s t test of precipitation rates.

RAMS-simulated total precipitation (CLAIRUN) was compared to observation to assess the model performance (Fig. 4, bottom panels). The Tropical Rainfall Measuring Mission (TRMM; Kummerow et al. 2000) satellite-retrieved precipitation was used in this study. The TRMM data have 3-h temporal resolution and 0.25° × 0.25° (∼27 km) spatial resolution. Figure 4 shows that RAMS captured the overall distribution: dry in the eastern domain and high rainfall over the Congo forest and around Lake Victoria. However, it underestimated precipitation in some areas, especially near the west and east boundaries.

Figure 4 (top panel) shows that RLAI and CLAI had substantial impacts on simulated precipitation. In this figure, dotted contour lines at a few scattered locations indicate statistically significant changes with a 90% confidence level. The spatial pattern of precipitation difference corresponds well to surface heat fluxes. Though the domain-averaged precipitation difference is small (0.88 mm), some regions experienced precipitation difference as large as 92.2 mm (Kenya). RAMS simulated much more precipitation over the Congo forest compared to the TRMM observation. Reduced latent heat flux and thus precipitation by using CLAI decreased this positive bias. Meanwhile, CLAIRUN produced considerably higher rainfall in the eastern domain (large red zone in Kenya and Tanzania; Fig. 4), which reduced the negative bias in the model. This increased rainfall is substantial for this relatively dry region.

2. Conclusions

Satellite-derived LAI data are increasingly being used in climate modeling studies and numerical weather predictions for improved characterization of land surface conditions. However, LAI in most land surface models is defined for the vegetated part only, while remote sensing LAI is for the total area (vegetated and nonvegetated). Using a regional model, this study found that this difference in definition caused large difference in LAI. As a result, the partition of surface available energy and simulated precipitation were affected substantially. The improper use of remote sensing LAI was found to increase the biases in simulated precipitation in RAMS. As the tendency of multidisciplinary research continues in the global environmental change study, this paper also demonstrated the importance of better communications among different fields.

Acknowledgments

The author would like to thank two anonymous reviewers for their valuable comments. The author also would like to thank Drs. Robert Dickinson, Lahouari Bounoua, and Gordon Bonan for their comments on the LAI definition. This research is supported by the Arts and Sciences Dean’s Incentive Grant, Oklahoma Sate University.

REFERENCES

REFERENCES
Betts
,
R. A.
,
2000
:
Offset of the potential carbon sink from boreal forestation by decreases in surface albedo.
Nature
,
408
,
187
190
.
Bounoua
,
L.
,
G. J.
Collatz
,
S. O.
Los
,
P. J.
Sellers
,
D. A.
Dazlich
,
C. J.
Tucker
, and
D. A.
Randall
,
2000
:
Sensitivity of climate to changes in NDVI.
J. Climate
,
13
,
2277
2292
.
Buermann
,
W.
,
J.
Dong
,
X.
Zeng
,
R. B.
Myneni
, and
R. E.
Dickinson
,
2001
:
Evaluation of the utility of satellite-based vegetation leaf area index data for climate simulations.
J. Climate
,
14
,
3536
3550
.
Castro
,
C. L.
,
R. A.
Pielke
Sr.
, and
G.
Leoncini
,
2005
:
Dynamical downscaling: Assessment of value retained and added using the Regional Atmospheric Modeling System (RAMS).
J. Geophys. Res.
,
110
,
D05108
.
doi:10.1029/2004JD004721
.
Chase
,
T. N.
,
R. A.
Pielke
,
T. G. F.
Kittel
,
R.
Nemani
, and
S. W.
Running
,
1996
:
Sensitivity of a general circulation model to global changes in leaf area index.
J. Geophys. Res.
,
101
, (
D3
).
7393
7408
.
Deardorff
,
J. W.
,
1978
:
Efficient prediction of ground surface temperature and moisture, with inclusion of layer of vegetation.
J. Geophys. Res.
,
83
, (
C4
).
1889
1903
.
Dickinson
,
F. E.
,
M.
Shaikh
,
R.
Bryant
, and
L.
Graumlich
,
1998
:
Interactive canopies for a climate model.
J. Climate
,
11
,
2823
2836
.
Feddema
,
J. J.
,
K. W.
Oleson
,
G. B.
Bonan
,
L. O.
Mearns
,
L. E.
Buja
,
G. A.
Meehl
, and
W. M.
Washington
,
2005
:
The importance of land-cover change in simulating future climates.
Science
,
310
,
1674
1678
.
Foley
,
J. A.
,
M. H.
Costa
,
C.
Delire
,
N.
Ramankutty
, and
P.
Snyder
,
2003
:
Green surprise? How terrestrial ecosystems could affect earth’s climate.
Front. Ecol. Environ.
,
1
,
38
44
.
Ge
,
J.
,
J.
Qi
,
B. M.
Lofgren
,
N.
Moore
,
N.
Torbick
, and
J. M.
Olson
,
2007
:
Impact of land use/cover classification accuracy on regional climate simulations.
J. Geophys. Res.
,
112
,
D05107
.
doi:10.1029/2006JD007404
.
Ge
,
J.
,
J.
Qi
, and
B. M.
Lofgren
,
2008
:
Use of vegetation properties from EOS observations for land-climate modeling in East Africa.
J. Geophys. Res.
,
113
,
D15101
.
doi:10.1029/2007JD009628
.
Huete
,
A.
,
K.
Didan
,
T.
Miura
,
E. P.
Rodriguez
,
X.
Gao
, and
L. G.
Ferreira
,
2002
:
Overview of the radiometric and biophysical performance of the MODIS vegetation indices.
Remote Sens. Environ.
,
83
,
195
213
.
Kalnay
,
E.
, and
Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
.
Kang
,
H-S.
,
Y.
Xue
, and
G. J.
Collatz
,
2007
:
Impacts assessment of satellite-derived leaf area index datasets using a general circulation model.
J. Climate
,
20
,
993
1015
.
Kummerow
,
C.
, and
Coauthors
,
2000
:
The status of the Tropical Rainfall Measuring Mission (TRMM) after two years in orbit.
J. Appl. Meteor.
,
39
,
1965
1982
.
Lawrence
,
P. J.
, and
T. N.
Chase
,
2007
:
Representing a new MODIS consistent land surface in the Community Land Model (CLM 3.0).
J. Geophys. Res.
,
112
,
G01023
.
doi:10.1029/2006JG000168
.
Lu
,
L.
, and
W. J.
Shuttleworth
,
2002
:
Incorporating NDVI-derived LAI into the climate version of RAMS and its impact on regional climate.
J. Hydrometeor.
,
3
,
347
362
.
Mayaux
,
P.
,
E.
Bartholome
,
S.
Fritz
, and
A.
Belward
,
2004
:
A new land cover map of Africa for the year 2000.
J. Biogeogr.
,
31
,
861
877
.
Myneni
,
R. B.
,
R. R.
Nemani
, and
S. W.
Running
,
1997
:
Estimation of global leaf area index and absorbed PAR using radiative transfer models.
IEEE Trans. Geosci. Remote Sens.
,
35
,
1380
1393
.
Oleson
,
K. W.
, and
G. B.
Bonan
,
2000
:
The effects of remotely sensed plant functional type and leaf area index on simulations of boreal forest surface fluxes by the NCAR Land Surface Model.
J. Hydrometeor.
,
1
,
431
446
.
Pielke
Sr.,
R. A.
,
1992
:
A comprehensive meteorological modeling system—RAMS.
Meteor. Atmos. Phys.
,
49
,
69
91
.
Pielke
Sr.,
R. A.
,
G.
Marland
,
R. A.
Betts
,
T. N.
Chase
,
J. L.
Eastman
,
D. S.
Niyogi
, and
S. W.
Running
,
2002
:
The influence of land-use change and landscape dynamics on the climate system: Relevance to climate-change policy beyond the radiative effect of greenhouse gases.
Philos. Trans. Roy. Soc. London
,
A360
,
1705
1719
.
doi:10.1098/rsta.2002.1027
.
Pitman
,
A. J.
,
2003
:
The evolution of, and revolution in, land surface schemes designed for climate models.
Int. J. Climatol.
,
23
,
479
510
.
doi:10.1002/joc.893
.
Sellers
,
P. J.
,
Y.
Mintz
,
Y. C.
Sud
, and
A.
Dalcher
,
1986
:
A Simple Biosphere Model (SiB) for use within general circulation models.
J. Atmos. Sci.
,
43
,
505
531
.
Sellers
,
P. J.
,
C. J.
Tucker
,
G. J.
Collatz
,
S. O.
Los
,
C. O.
Justice
,
D. A.
Dazlich
, and
D. A.
Randall
,
1994
:
A global 1° by ° NDVI data set for climate studies. Part 2: The generation of global fields of terrestrial biophysical parameters from the NDVI.
Int. J. Remote Sens.
,
15
,
3519
3545
.
Turner
II,
B. L.
,
E. F.
Lambin
, and
A.
Beenberg
,
2007
:
The emergence of land change science for global environmental change and sustainability.
Proc. Natl. Acad. Sci. USA
,
104
,
20666
20671
.
van den Hurk
,
B. J. J. M.
,
P.
Viterbo
, and
S. O.
Los
,
2003
:
Impact of leaf area index seasonality on the annual land surface evaporation in a global circulation model.
J. Geophys. Res.
,
108
,
4191
.
doi:10.1029/2002JD002846
.
Walko
,
R. L.
, and
Coauthors
,
2000
:
Coupled atmosphere–biophysics–hydrology models for environmental modeling.
J. Appl. Meteor.
,
39
,
931
944
.
Xue
,
Y.
,
1997
:
Biosphere feedback on regional climate in tropical north Africa.
Quart. J. Roy. Meteor. Soc.
,
123
,
1483
1515
.
doi:10.1002/qj.49712354203
.

Footnotes

Corresponding author address: Jianjun Ge, Department of Geography, Oklahoma State University, Stillwater, OK 74078. Email: jianjun.ge@okstate.edu