(Ruimy et al., 1999; Xiao et al., 2004). The LUE was calculated by the following equation of GPP=LUE*fAPAR*PAR, GPP were from EC measurements, fAPAR from in situ LAI, and PAR from in situ meteorological measurements. We think that it was not appropriate to use MODIS product for calculation.
7. Section 3.2. \Again,
the authors do not provide how \generate
a daily comparison with the satellite. Have the authors considered that satellite overpass time varies day
to day across time zones (hourly intervals)?
Response: we added the data processing information in section 2.3 and the daytime here is the MODIS
satellite overpass time.
8. There needs to be a minimal level of aggregation of the fluxes in order to be able to compare with
satellite data, and the authors need to provide this information.
Response: we consider it is necessary to give more detail information about the data used. In section
2.3, with the MODIS overpass time, we used 5 readings of NEE around and 10 readings of PAR for
aggregation and averaged in later analysis.
9. Section 3.3. How were LUE values computed and derived? is this information provided? Response: see our explanations with concerns of No.6.
10. Section 3.4.1. \largely
affected by background information..\affected
by different soils, but the authors did not establish that (1) there are soil variations of concern within this
study and (2) and that these were responsible for the poorer NDVI results.
11. Same applies to: \and
sky conditions than other indices.\in this
study and are merely guessing why the indices yield different results. As it was not objective of this
study, such interpretations should be removed from Results; although one is free to conjecture such
ideas in the Discussion section (so move such comments to discussion section).
Response: we checked throughout the manuscript and removed these parts into a new discussion part.
12. Conclusions section: \this paper, our method worked well for the wheat that was a relatively
homogenous canopy,\ Response: We are sorry for this mistake.
13. There should be some error bars in Figs 3-5. Response: we added the error bars as suggested.
14. It was annoying that no page numbers were used. Response: page numbers were added in this revision. Reviewer #2 (Comments):
This paper investigates the relationship between gross primary production (GPP) and four vegetation
indices (NDVI, EVI, MSAVI, and WDVI), using data from a maize site. Authors aim to seek a simple
relationship between GPP and vegetation indices, which is an important effort in linking satellite observations with CO2 flux tower measurement. However, I have a number of concerns on the manuscript.
1. English writing needs substantial improvement, and I would suggest that authors seek a help from a
native English speaker to work on it. The manuscript does not have page number and line number, which makes it difficult to write comments for the manuscript. The reference citation in the text does
not follow JGR format.
Response: we got help from a native speaker to proofread the paper and the reference citations were
changed to follow the JGR format.
2. Page 2 Introduction: 1st paragraph needs to be re-written and expanded to include more scientific
justification on their study of GPP.
Response: we added more information about the GPP definition and current mechanism of GPP estimation based on LUE model.
3. Page 4 literature review on the VPM model. Note that Yan et al., 2009 used the VPM model to simulate
GPP of both winter wheat and maize crop in a study site in North China. The paper might be of interest to the authors.
Yan, H., Fu, Y., Xiao, X., Huang, H., He, H. and Yu, G., 2009, Modeling gross primary productivity of
winter wheat and maize double-cropping system, using MODIS time series imagery and CO2 eddy flux
data, Agriculture, Ecosystems and Environment, 129(4): 391-400.
Response: yes, we added this reference to give a better introduction of both VPM model and GPP estimation.
4. Page 5. It mentioned \ Response: We added three VIs (PRI, NDSI and MTCI) as examples for LUE estimation to specify the use of VIs.
5. Page 5 last paragraph, \it
in a new paragraph.
Response: we followed this suggestion in the manuscript.
6. Page 6. Section 2.1. It is too short, and needs to expand for providing more info about the site, particularly vegetation and soils.
Response: additional information were added to better description of this site.
7. Page 6, Section 2.2, how frequently does LAI is measured? Response: The LAI was measured everyday during the experiment time.
8. Page 7, section 2.4. Is the daily MODIS data used in the study? Is the MODIS onboard Terra or the
MODIS onboard Aqua satellite used? When 3x3 pixels (1.5km) are used? What is the fraction of maize
within the 1.5 km radius?
Response: we used the MODIS/Terra daily surface reflectance data in this paper. The 3x3 pixels were
used for calculation of reflectance covering the EC sites. In the maize site, the fraction of maize within t
he 1.5 km radius would be higher than 95% because only small ridges were inside. For forest and grassland sites, the fraction could be almost 100%.
9. Section 2.5, page 7 - Page 8, it needs to expand and provide detailed description on how LUE was
calculated. Maize is a summer crop, why authors used data from 11/1/2007 to7/12/2008 to derive a
relationship between LAI and fapar?
Response: in section 2.5, we provided a more detail information of LUE and fAPAR determination. Furthermore, we also consider it was not appropriate to use MODIS LAI/fAPAR product to derive the relationship. Instead, we used a method suggested by reviewer1 to use the in situ measured LAI
for the calculation of fAPAR with equation of fAPAR=0.95(1-Exp (-0.5LAI)). This method demonstrated
to be workable with other publications (Ruimy et al., 1999; Xiao et al., 2004). The LAI used was data
in the same time of other observations.
10. Page 8, Section 2.6, authors need to specify what indices were used the work of Inoue et al. (2008).
Response: some VIs were added in the part to avoid confusion.
11. Page 10, Section 3.1, why only data (NEE and T) from 22:00pm to 3:00am next day are used in
calculated? The night time definition needs to revisit and authors should use appropriate definition of
night time? In short, authors need to provide much better and consistent way to describe how to partition NEE data into GPP and ecosystem respiration.