Analysis of satellite observed SiF and fPAR during the Russian
heatwave of summer 2010
A study of the added value of space borne SiF measurements
Lisa Mulder
10366431
July 11, 2017
Report Bachelor Project Physics and Astronomy, size 12 EC
Conducted between 03-03-2017 and 13-07-2017
Netherlands Institute for Space Research, SRON
University of Amsterdam, Faculty of Science
Supervisor: Dr. Ir. S. Houweling
Second Assessor: Prof. Dr. E.A.A. Aben
Natuurkunde practicum, Amsterdam
Contents
1
Populaire samenvatting
4
2
Introduction
5
3
The Russian drought
9
4
Satellite datasets
12
5
Method
13
6
Results
16
7
Discussion
27
8
Conclusion
28
9
Bibliography
28
Abstract
SiF (Solar induced Fluorescence) measurements, provided by the GOME-2 spectrome-ter on board of the MetOp-A satellite, are shown to have an added value over fPAR (frac-tion of Absorbed Photosynthetically Active Radia(frac-tion), provided by Moderate Resolu(frac-tion Imaging Spectroradiometer (MODIS) instrument onboard the Aqua and Terra satellite, in terms of the rate to measure developments in the atmosphere, such as droughts. fPAR is a measure of the greenness of the vegetation, which correlates to photosynthesis. SiF has found to be a bridge between photosynthesis capacity and rate, as it provides more direct information about the photosynthesis process. The possible added value, temporal response, of SiF was examined through an anomaly, the summer heatwave of 2010. The heatwave caused severe changes in the conditions for vegetation, high temperatures and the forest fires led to air pollutants. As examined by Guerlet et al. (2013) the difference in CO2fluxes between 2009 and 2010 reached their peak in June, July and August. fPAR
shows no signal of the drought, except for lower values in 2010 at the end of August. The SiF values found show an effect of the drought in June and July. So SiF measure-ments give a more accurate representation of the CO2 uptake from the atmosphere, by
1
Populaire samenvatting
De toenemende opwarming van de aarde heeft er voor gezorgd dat er meer behoefte is aan representatieve informatie over wat er op aarde gebeurt. Al een langere tijd worden satellieten gebruikt om ons informatie te geven over de atmosfeer. Een aspect die kan worden onderzocht is de opname van CO2, oftewel de photosynthese. Een nieuw satelliet
product hiervoor is SiF (Solar induced Fluorescence), deze methode is erop gebaseerd dat chlorophyll een bepaald signaal afgeeft tijdens zijn aangeslagen toestand, het fotosynthese proces. Wetenschappers denken dat SiF een betere weergave geeft van de CO2 opname,
aangezien het een sterkere correlatie heeft met fotosynthese dan andere producten. Een voorbeeld van een ander product is fPAR (fraction Absorbed Photosynthetically Active Radiation), dit meet de groenheid van vegetatie en leidt daaruit de fotosynthese af. Een van de mogelijke verbeteringen ten opzichte van fPAR, is het reactievermogen van SiF op anomalien in de atmosfeer, zoals droogte of wijzigingen van de samenstelling in de atmos-feer. In dit onderzoek werd er gekeken naar de hittegolf in 2010. Deze hittegolf zorgde voor extreme temperaturen en bosbranden in de regio Moskou. Extreme temperaturen en de emissies van de branden zorgen voor andere condities voor de vegetatie, wat zijn weergave heeft op de mate van fotosynthese. In dit onderzoek werd er gebruik gemaakt van een onderzoek, gedaan door Guerlet et al. (2013), zij onderzochten de verschillen in CO2 fluxes tussen de jaren 2009 en 2010. Bij het kijken naar de verschillen in fPAR en
SiF waardes tussen de jaren 2009 en 2010, werden de verwachtingen bevestigd. fPAR gaf een verlaat teken van de hittegolf in Augustus, maar gaf bijna geen verschillen aan in de maanden Juni en Juli. SiF daarentegen gaf voor elke maand verschillen aan, tussen de jaren 2009 en 2010. Dit betekent dat SiF een sneller reactievermogen heeft dan fPAR en dus een betere weergave van de CO2 opname in de atmosfeer geeft.
Figure 1: Weergave van de verschillen in fPAR values tussen 2009 en 2010. Rode vierkant
geeft het sterk benvloede gebied rond Moskou weer.
2
Introduction
The global mean surface temperature has been increasing since the 19th century, 1 degree since 18501. Climate models predict that the global annual surface temperature will in-crease between 1 and 3.5◦C by 2100 (Hughes, 2000), due to the increase of atmospheric CO2 and other greenhouse gases. The growth of CO2 concentrations has severe
conse-quences for the climate and the flora and fauna on earth (figure 2). Recent changes to the physical features of the earth, such as the rise of sea levels, the decrease in glaciers and sea ice and changes in weather extremes have sparked agreements for the regulation of CO2 emissions, e.g. the COP21 climate agreement.
Figure 2: The direct and indirect effects of CO
2and other greenhouse gases on climate,
physiology, phenology and distributions of species (Hughes, 2000).
The mixing ratio of carbon dioxide in the atmosphere and its yearly increase are determined by three main processes. Anthropogenic emissions (mainly from fossil use), uptake and release by vegetation (photosynthesis and respiration) and absorption by the oceans (figure 3).
The increase of anthropogenic emissions is partially compensated by absorption in the oceans and the biosphere on land. The compensation is caused by the imbalance of the Henry balance, the absorption restores the solubility equilibrium (Dacey et al., 1987). For vegetation its more complicated. CO2 is a nutrient for vegetation, so the increase
causes growth of trees and plants, which leads to an increase in CO2 uptake. There is
an uncertainty, if vegetation will continue with a netto carbon dioxide uptake. There is a possibility, that climate changes will make it unfavorable for vegetation to absorb CO2.
An important aspect is the climate sensitivity of vegetation, this is still very uncertain. The aim of this research is to measure the reaction of vegetation on climate disturbances, through satellite measurements.
Figure 3: The global carbon cycle with the contributions of humankind (red) and nature
(yellow), in billions of tons per year. The storage of carbon is indicated in white.
Vegetation-atmosphere interactions such as photosynthesis can dampen or amplify anthropogenic climate change (Bonan, 2008). As mentioned before, now it dampens the climate change in the future it can be amplified. Vegetation contributes to the amount of carbon in our atmosphere by plant respiration and decreases it through photosynthesis.
Satellite measurements have been used to determine vegetation activity, including the amount of carbon stored through photosynthesis. This is quantified by the Gross Primary Productivity (GPP), which is a measure of the rate at which photosynthesis occurs.
GPP can be inferred from satellite measurements for example, NDVI (normalized dif-ference vegetation index) or fPAR (fraction of Photosynthetically Active Radiation).These methods determine the greenness of vegetation, which is related to the amount of leaf chlorophyll. It quantifies the capacity to store carbon, however this says little about the actual photosynthesis rate.
The recently developed Solar Induced fluorescence (SiF) data product from satellites, is expected to bridge the gap between photosynthesis capacity and rate, as it provides more direct information about the photosynthesis process. The SiF method is, based on the characteristic fluorescence in the red spectral region during the excited state of chlorophyll. Chlorophyll is essential during the process of photosynthesis, because it collects the required energy by absorbing sunlight. However, a fraction of the energy is not used for the synthesis of sugars, but escapes als fluorescent light. This allows the plant to control the amount of energy, and thereby avoid damage to the photosystem. Although fluorescence generally increases with increasing availability of light and photosynthesis, it is also a signal of plant stress. This happens, for example, when less CO2 is available
for photosynthesis because the stomata are closed to avoid evaporation when the water availability is low. Although this complicates the relation between SiF and photosynthesis, the analyses that have been published so far nevertheless point to fairly easy relationships between SiF and GPP (Frankenberg et al., 2014).
Carbon uptake has substantial spatiotemporal variability, through the relation between carbon uptake and vegetation productivity (Chahine et al., 2008).
Vegetation productivity is dependent on the availability of several resources, such as CO2, other nutrients (e.g. N, P), water and physical conditions such as light and
temperature, such as droughts, floods and heat waves influence the productivity and the rate of photosynthesis (Berry & Downton,1982). With the availability of SiF data from satellites, the question arises what can be learned about vegetation productivity in addition to the information that is provided already by established techniques such as NDVI. In a previous study (Aengenheyster, 2016), high correlations were found between SiF and NDVI, confirming the value of SiF as a measure of GPP. However, from this study to added value of SiF remains unclear. This outcome could result from the monthly time resolution that was used (Aengenheyster, 2016). The difference between vegetation greenness and actual photosynthesis may be larger at shorter timescales, as photosynthesis is expected to change much quicker than the colour of the leaves.
analyzing sub-monthly time variations in vegetation productivity. This is achieved by replacing the monthly NDVI data from AVHRR used in Aengenheyster (2016), with MODIS fPAR(fraction of Absorbed Photosynthetically Active Radiation) improving the time resolution to 8 days. fPAR is a measure of the greenness of the vegetation, similar to NDVI except that it makes use of information from a broader wavelength range (0.4 -0.7 µm). To make use of this improved dataset, the analysis in this thesis concentrates on the summer heatwave in Eurasia of 2010.
This research focuses on comparing the signal of the summer heatwave 2010, in the satellite products SiF and fPAR (Fraction of Absorbed Photosynthetically Active Radia-tion), to answer the question: Does SiF (Solar induced Fluorescence) have an added value above fPAR (fraction of Absorbed Photosynthetically Active Radiation) for the rendition of photosynthesis in terms of the rapidity of the reaction to the summer heatwave 2010 on the Northern Hemisphere?
3
The Russian drought
The spatio-temporal evolution of the summer heatwave in 2010, has been examined by Barriopedro et al. (2011). The heatwave was centered over Western Russia, which experi-enced the longest droughts and the largest temperature deviations. Historical temperature records were broken from the beginning of July until the end of August 2010 (see figure 3). The 2010 summer heatwave was unique regarding its temporal amplitude and spatial extent (2000 km2). In July and August temperatures were found +16 and +18 (◦C)
av-eraged above normal with a time frame of 15 days (Joiner et al., 2014). Also a relatively low humidity was found (9 and 25 %) from mid June to mid-August 2010 (Witte et al, 2011).
Figure 4: Spatio-temporal evolution of the summer heatwave of 2010. Maximum temperature
anomalies for 7-day, 15-day, 31-day and 81-day periods. The contour lines connect location
with the same anomaly divided by the corresponding standard deviation (SD) of all summer
days of the reference period (Barriopedro et al., 2011).
In July and August the strong heat wave caused severe wildfires in the western part of Russia, leading to large emissions of aerosols and trace gases, such as hydrocarbons, nitrogen oxides and CO. On the 29th of July the fires started and on the second of August the wildfires spread to an area of 5000 km2 and caused carbon dioxide levels 3.6 times
above the allowed numbers (Huijnen et al., 2012).
This short-term anomaly has multiple consequences for the photosynthesis in the re-gion of Western Russia. The temperature range suitable for photosynthesis is approxi-mately between 10 and 35◦C. Above the level of tolerance, there will be irreversible loss in photosynthesis capacity of vegetation, see figure 5 (Berry & Downton, 1982, pg 294).
Figure 5: Temperature dependence of net CO
2uptake in the range of 10 to 50
◦C (Berry &
Downton, 1982, pg 296).
Drought and high temperatures often occur simultaneously, in Shah & Paulson (2003) they examined the impact of these variables on photosynthesis. The variable drought (plant water potential of .0 to 2.4 MPa) has a negative effect on photosynthesis, as it de-creases viable leaf area, as well as photosynthesis. In combination with high temperatures, the effects of drought are more severe. The productivity of the vegetation is compromised and a loss in photosynthesis activity is found.
The respiration of vegetation is affected by air pollutants, directly and indirectly. Despite the elevated levels of CO2during forest fires, smoke-haze decreases photosynthesis
activity within vegetation (Davies & Unam, 1999). An effect is the reduction of PAR (photosynthetically active radiation) levels.
Davies & Unam (1999) found a decrease of 45-92% in PAR levels, compared to a clear day. The increase of aerosol and atmospheric pollutant levels are the cause of another negative effect on photosynthesis. During the heatwave, two processes cause the elevated CO2 levels: emissions from the fire and reduced biospheric uptake due to drought and
smoke. Guerlet et al. (2013) examined these separate contributions. The contribution of fire emissions to the total CO2 emission cant be quantified, however contribution of fires
could be significantly higher than expected (see figure 6).
Figure 6: The difference in CO
2-emission between the years 2009 and 2010, in
monthP gC. The
green diamonds represent the assimilation of flask data and the red squares the joint inversion
of GOSAT and surface data. The top image uses prior biospheric emissions from the CASA
model, using GFED- derived burned area (dashed dark red line) and prior fire emissions
(dashed red line). Bottom image considers no prior IAVs. The blue shaded area are the fire
estimates based on ALANIS (Guerlet et al, 2013).
The heavy smoke emission during the heatwave was examined by Krol et al., 2013. The smoke probably reduced the shortwave radiation, which caused substantial heating of the atmosphere by radiation absorption (Krol et al, 2013).
4
Satellite datasets
The method of Solar Induced Fluorescence (SiF) is a characteristic fluorescence of chloro-phyll in the red spectral region, during the excited state (caused by laser beam or any light source, with a wavelength in the blue or orange spectral region). The correlation be-tween chlorophyll fluorescence and photosynthesis has been proven in previous researches (Maier, 2002). As mentioned before, photosynthesis makes use of solar irradiation (380-750 nm). This is the absorbed photosynthetic active radiation (PAR) that plants use to convert and store solar energy in chemically bound energy (Maier, 2002). The retrieval of SiF from satellite observed spectra of reflected solar radiation is achieved using the weak signal measured at the wave lengths Fraunhofer lines (Joiner et al. 2011). The measurement takes place in the dip of a Fraunhofer line to optimize the ratio between the solar reflectance signal and weak fluorescence signal of vegetation (Smorenburg et al., 2002).
fPAR represents the energy absorption capacity of vegetation, fPAR is defined as the fraction of PAR, that is absorbed energy by vegetation. fPAR depends on reflectance and transmission spectra of the canopy, the incident radiation field and the reflectance of the background (Ganguly et al.,2014, pg 44) .
The SiF dataset that has been used in this study was retrieved from the optical spectrometer GOME-2 onboard the MetOp-A satellite, which collects spectra at near global coverage daily. The data was processed with an algorithm developed at the KNMI by Tommy van Leth and applied to the full measurement dataset of GOME-2 by Maurits Kooreman. The data are retrieved at a daily temporal granularity and a 80 x 40 km footprint size. (Aengenheyster, 2016).
The fPAR dataset that has been used was derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard the Aqua and Terra satellite. The MODIS dataset is made available at the temporal granularity of eight days and a pixel size of 500 meters (Myneni et al, 2002).
MODIS data is filed in a sinusoidal tile grid. The projection is pseudo cylindrical, where all the parallels and central meridian are straight (see figure 7). The meridians are determined by a sine function. The amplitudes of the sine function have a positive correlation with the distance to the central meridian. The sinusoidal projection maintains equal area at all latitudes, but creates distortion of the shapes and directions2. MODIS
data is available online3.
2http://desktop.arcgis.com/en/arcmap/10.3/guide-books/map-projections/sinusoidal.htm 3https://reverb.echo.nasa.gov/reverb
Figure 7: Sinusoidal projection with Tissots indicatrix of deformation (orange circles).
5
Method
This thesis exams tiles: 19-02, 19-03, 20-02 and 20-04 (see figure 7), corresponding to the area which was affected the most by the heatwave (Wolfe, 2017).
Figure 8: Overview of the MODIS tile system (Wolfe, 2017).
In this study, the MODIS data has been reprojected to a equirectangular projection using the MODIS reprojection tool (MRT). MRT is a software program made available to facilitate the use of MODIS data. The equirectangular projection converts the globe in Cartesian (equal) rectangles in a coordinate system of degrees of latitude and longitude. Usually the equator is chosen as the central parallel. The horizontal component is the longitude and the vertical component is the latitude4(see figure 8).
Figure 9: Equirectangular projection with Tissots indicatrix of deformation.
The MODIS data product that was used is the MOD15A2H V006 dataset, which in-clude Leaf Area Index (LAI) and fPAR. MRT supports three optional output file formats: hdf-eos, raw binary and geotiff. This thesis used hdf-eos for further processing. The as-pect Resampling Type has as well three options: nearest neighbour, bilinear and cubic convolution. Nearest neighbour is for discrete and continuous data. This thesis used Nearest neighbour for the possibility of discrete data (see figure 9).
MRT needs additional information about the data for the correct change in projections. For the equirectangular projection MRT needs information about the Sphere, CentMer, TrueScale, FE and FN.
Sphere is the radius of the reference sphere, which has been set at the Earths radius (6370997 meters). CentMer represents the longitude of the central meridian. TrueScale is the latitude of the lines of true scale, with the sinusoidal mapping all parallels are standard (lines of true scale). FE is false-easting, this needs to be in the same units as the semi-major axis. FN is the same as FE, except that FN is false-northing . The values used in this thesis are represented in figure 11 (Land Processes DAAC USGS Earth Resources Observation and Science (EROS) Center, 2011).
6
Results
As previously mentioned, Guerlet et al. (2013) found a reduced carbon uptake during the Russian heatwave 2010 in the period of June to, August 2010. Their results showed a large difference in satellite inferred CO2 fluxes into the atmosphere between June 2009
and June 2010 (see figure 6). The corresponding fPAR values for the year 2009 and 2010 are shown in figures 12 and 13. The fPAR values are in percentages with a fill value of 245 indicating missing data. The red square highlights the heavily affected area around Moscow. Barriopedro et al. (2011) found the highest temperature deviations in this area. Witte et al. (2011) measured the highest fire counts of Europe in this area.
Figure 14 shows the difference in fPAR between 2009 and 2010, without fill values. Fill values complicated the analysis between the years 2009 and 2010. In the highlighted area values alternate between small positive and negative differences. Positive differences denote higher fPAR values in 2010. The differences in fPAR values around Moscow for June between the years are small. To the North there is a region with differences around 40/60 and - 40/60. Therefore, for this region large positive and negative differences were found between 2009 and 2010.
Figure 12: fPAR values for June 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 13: fPAR values for June 2010. Red boxed area is the heavily affected area around
Moscow.
Figure 14: Difference in June fPAR between the years 2009 and 2010, without fill values. Red
boxed area is the heavily affected area around Moscow.
As seen in Guerlet et al.(2013) the satallite inferred of CO2flux anomalies are largest
in July and August. So the expectation is a great difference between the fPAR values of 2009 and 2010. Figure 15 and 16 represent the fPAR values for July in 2009 and 2010.
Figure 15: fPAR values for July 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 16: fPAR values for July 2010. Red boxed area is the heavily affected area around
Moscow.
Figure 17 shows the differences in fPAR values for July, without fill values. The highlighted area and the rest of the area show almost no differences between 2009 and 2010, except an area North of Moscow, where differences are found of 60/80 percent.
Figure 17: Difference in fPAR values for July between the years 2009 and 2010, without fill
values. Red boxed area is the heavily affected area around Moscow.
In August large differences were found in West Eurasia between 2009 and 2010 (Guerlet et al., 2013).The fPAR values for August 2009 and 2010 are shown in figure 18 and 19.
In August large negative differences were found between 2009 and 2010, as seen in figure 20, outside the highlighted area and the highly affected area around Moscow.
Figure 18: fPAR values August 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 19: fPAR values August 2010. Red boxed area is the heavily affected area around
Moscow.
Figure 20: Difference in August fPAR between the years 2009 and 2010, without fill values.
Red boxed area is the heavily affected area around Moscow.
The SiF values for June 2009 and 2010 are seen in figure 21 and 22, with unit W/sr/m2/nm. The highly affected area is once again shown in the red square. The figures showing the SiF differences between the years 2009 and 2010, as well as the figures for the separate months are exclusive fill values.
Figure 23 confirms the outcome from Guerlet et al. (2013), there are elevated SiF values for the highlighted area. In general theres a positive difference between the SiF values of 2009 and 2010, except for an area Western from Moscow. Positive values denote higher SiF values in 2010.
Figure 21: SiF values for June 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 22: SiF values for June 2010. Red boxed area is the heavily affected area around
Moscow.
Figure 23: Difference in June SiF between the years 2009 and 2010. Red boxed area is the
heavily affected area around Moscow.
The SiF values for July are shown in figure 24 and 25.
Figure 24: SiF values for July 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 25: SiF values for July 2010. Red boxed area is the heavily affected area around
Moscow.
There are almost no differences between the years 2009 and 2010 in July. Although, somewhat complicated by the color scale, most of the differences are around zero W/sr/m2/nm.
Figure 26: Difference in July SiF between the years 2009 and 2010. Red boxed area is the
heavily affected area around Moscow.
Figures 27 and 28 show the SiF values for August 2009 and 2010.
Figure 27: SiF values August 2009. Red boxed area is the heavily affected area around
Moscow.
Figure 28: SiF values for August 2010. Red boxed area is the heavily affected area around
Moscow.
Just outside the highlighted area great differences are found between 2009 and 2010 (figure 29). There are as well regions with negative differences. In the highlighted area there are positive and negative differences.
Figure 29: Difference in July SiF between the years 2009 and 2010. Red boxed area is the
heavily affected area around Moscow.
For a clearer view of the SiF and fPAR values a time analysis have been done for the months June, July and August, with time intervals of 8 days. The years 2009 and 2010 are compared for a clearer view of the differences. For a time series representing only the changes in the highly affected area, data points in the range of latitude (54.29 - 58.42) and longitude (33.06 - 41.11), were used (showed as the red square in previous figures). For two time intervals in June 2010, no data was retrieved. When examining figure 30 no significant signal of the impact of the drought is visible, except for the end of August. For the last two intervals, fPAR of the year 2010 is significantly lower, than fPAR of 2009.
Figure 30: Temporal variations in fPAR and their standard deviations for the months June,
July and August of the years 2009 and 2010, in percent. For the highly affected area around
Moscow.
The time analysis for SiF shows a more coherent and less noisy signal. Figure 31 shows a clear difference between 2009 and 2010 for June. For August as well, but smaller differ-ences were found between 2009 and 2010. For June and August SiF values are elevated in 2010 compared to 2009. In July small differences were found with overlapping standard deviations, therefore no significant signal of the impact of the drought can be concluded.
Figure 31: Temporal variations in SiF and their standard deviations for the months June,
July and August of the years 2009 and 2010, in W/sr/m
2/nm. For the highly affected area
around Moscow..
7
Discussion
The MetOp-A satellite has a daily coverage. Because of the differences in temporal gran-ularity between MetOp-A and the Terra satellite, the SiF values were averaged every 8 days. By averaging the values of SiF, significance is lost. If products like fPAR and NDVI (Normalized Difference Vegetation Index) get data with a daily granularity, further re-search between SiF and products, like NDVI and fPAR, is recommended. Now you’re still limited to a 8-days temporal granularity, however a daily granularity may be unachievable because of the time needed to measure the whole earth.
In this research, we compared the fPAR and SiF values for the years 2009 and 2010. By comparing only two years, an uncertainty is introduced, because of possible anomalies in 2009. The recommendation for further research is to extend analysis to more years of data comparing 2010 with an average of SiF or fPAR over previous years.
As mentioned before the Terra and the MetOp-A satellite differ in spatial resolution, this could affect the results. With SiF and fPAR values, missing data was filled with a fill value. Ideally when examining data, you only want to use data with a value for every time interval. By looking at the area around Moscow for the temporal variations, influences of data points with great distances from Moscow were minimized. However, by reducing the spatial resolution data points are also reduced.
8
Conclusion
The aim of this research was to investigate the potential added value, of SiF concerning the temporal response of the vegetation to the Russian heatwave in the summer of 2010. This was done by examining SiF and fPAR for the months June, July and August of the years 2009 and 2010 and comparing these values with the expected values from previous studies, such as Guerlet et al. 2013 (figure 6).
For June the results were not in line with the expectations, found differences in fPAR values were in a region North of Moscow, rather than in the heavily affected area around Moscow. The highlighted area shows almost no differences. In July there is also an area with large differences between 2009 and 2010 North of Moscow. Once again it was hard to see a difference between 2009 and 2010 in the highlighted area. In August, large negative differences were found for the highlighted area and areas around the highlighted area. Negative differences are caused by higher fPAR in 2009, compared to 2010. This is in line with the expected differences, this could confirm the hypothesis of a delay in the signal of fPAR. A confirmation of the delay can possibly be found by examining September for the years 2009 and 2010. If there is an anomaly in fPAR, a delayed signal of the summer heatwave can be the cause.
The time analysis of regionally analysed fPAR (figure 30) did not provide further insides, except the difference in fPAR for late August. As mentioned before, the sudden differences in fPAR between 2009 and 2010 for August, can be a delayed signal of the drought.
Clear differences were found for the months June and August, however almost no differences were found for July, the time analysis confirms this. However, the positive differences mean higher SiF in 2010, than in 2009, which is not in line with expectations. An explanation can be: higher photosynthesis at the beginning of the drought, because of e.g. more sunlight. As the drought progresses, the signal of SiF drops faster, than in 2009. The time analysis agrees with this explanation, however the elevated SiF values in 2010 compared to 2009, in late August, was not expected.
In conclusion, SiF measurements give a more accurate representation of the CO2
uptake from the atmosphere, by vegetation. fPAR showed no significant signal of the drought, until August. The SiF differences show effects of the drought, since June.
9
Bibliography
1. Aengenheyster, M. (2016). Making, Analyzing and Interpreting Observations Project: Soil Moisture and Vegetation from Satellite Data (master thesis).
2. Barriopedro, D., Fischer, E. M., Luterbacher, J., Trigo, R. M., & Garca-Herrera, R. (2011). The hot summer of 2010: redrawing the temperature record map of Europe. Science, 332(6026), 220-224.
3. Berry, J. A., & Downton, W. J. S. (1982). Environmental regulation of photosyn-thesis. Photosynthesis, 2, 263-343.
4. Bonan, G. B. (2008). Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science, 320 (5882), 1444-1449.
5. Chahine, M. T., Chen, L., Dimotakis, P., Jiang, X., Li, Q., Olsen, E. T., ... & Yung, Y. L. (2008). Satellite remote sounding of midtropospheric CO2. Geophysical
Research Letters, 35 (17).
6. Dacey, J. W., Wakeham, S. G., & Howes, B. L. (1984). Henry’s law constants for dimethylsulfide in freshwater and seawater. Geophysical Research Letters, 11 (10), 991-994.
7. Davies, S. J., & Unam, L. (1999). Smoke-haze from the 1997 Indonesian forest fires: effects on pollution levels, local climate, atmospheric CO 2 concentrations, and tree photosynthesis. Forest Ecology and Management, 124 (2), 137-144.
8. Frankenberg, C., O’Dell, C., Berry, J., Guanter, L., Joiner, J., Khler, P., ... & Taylor, T. E. (2014). Prospects for chlorophyll fluorescence remote sensing from the Orbiting Carbon Observatory-2. Remote Sensing of Environment, 147, 1-12. 9. Ganguly, S., Nemani, R. R., Baret, F., Bi, J., Weiss, M., Zhang, G., ... & Singh, K.
(2014). Green leaf area and fraction of photosynthetically active radiation absorbed by vegetation. Biophysical Applications of Satellite Remote Sensing (pp. 43-61). Springer Berlin Heidelberg.
10. Guerlet, S., Basu, S., Butz, A., Krol, M., Hahne, P., Houweling, S., ... & Aben, I. (2013). Reduced carbon uptake during the 2010 Northern Hemisphere summer from GOSAT. Geophysical Research Letters, 40 (10), 2378-2383.
11. Hughes, L. (2000). Biological consequences of global warming: is the signal already apparent?. Trends in ecology evolution, 15 (2), 56-61.
12. Huijnen, V., Flemming, J., Kaiser, J. W., Inness, A., Leito, J., Heil, A., ... & Dufour, G. (2012). Hindcast experiments of tropospheric composition during the summer 2010 fires over western Russia. Atmospheric Chemistry and Physics, 12 (9), 4341-4364.
13. Joiner, J., Yoshida, Y., Guanter, L., Lindstrot, R., Voigt, M., Jung, M., ... & Frankenberg, C. (2014, April). New Measurements of Chlorophyll Fluorescence with Gome-2 and Comparisons with the Seasonal Cycle of GPP from Flux Towers. Proceedings of the 5th INTErnational Workshop on Remote Sensing of Vegetation Fluorescence, Paris, France (pp. 22-24).
14. Joiner, J., Yoshida, Y., Vasilkov, A. P., & Middleton, E. M. (2011). First observa-tions of global and seasonal terrestrial chlorophyll fluorescence from space. Biogeo-sciences, 8 (3), 637-651.
15. Krol, M., Peters, W., Hooghiemstra, P., George, M., Clerbaux, C., Hurtmans, D., ... & Kaiser, J. W. (2013). How much CO was emitted by the 2010 fires around Moscow?. Atmospheric Chemistry and Physics, 13 (9), 4737-4747.
16. Land Processes DAAC USGS Earth Resources Observation and Science (EROS) Center (2011). MODIS reprojection tool users manual. Retrieved at: 21 juni 2017, at https : //lpdaac.usgs.gov/sites/def ault/f iles/public/mrt41usermanual032811.pdf
17. Maier, S. W. (2002). Remote sensing and modelling of solar induced fluorescence. Remote Sensing of Solar-Induced Vegetation (Vol. 527).
18. Myneni, R. B., Hoffman, S., Knyazikhin, Y., Privette, J. L., Glassy, J., Tian, Y., ... & Lotsch, A. (2002). Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote sensing of environment, 83 (1), 214-231. 19. Shah, N. H., & Paulsen, G. M. (2003). Interaction of drought and high temperature
on photosynthesis and grain-filling of wheat. Plant and Soil, 257 (1), 219-226. 20. Smorenburg, K., Courreges-Lacoste, G. B., Berger, M., Buschman, C., Del Bello,
U., Langsdorf, G., ... & Visser, H. (2002, January). Remote sensing of solar-induced fluorescence of vegetation. International Symposium on Remote Sensing (pp. 178-190). International Society for Optics and Photonics.
21. Witte, J. C., Douglass, A. R., Silva, A. D., Torres, O., Levy, R., & Duncan, B. N. (2011). NASA A-Train and Terra observations of the 2010 Russian wildfires. Atmospheric Chemistry and Physics, 11 (17), 9287-9301.
22. Wolfe, R. (2017). Sinusoidal Tile Grid. Retrieved 20 June 2017, at https : //modis− land.gsf c.nasa.gov/M ODLAN Dgrid.html
