1 Dazilkree

Evi 3702 Assignment 1 2013


Plant phenology is usually defined as the timing of periodic events in the life cycle of plants (Piao et al., 2015; White et al., 2009). For any given biome, phenology events, such as bud-burst, flowering, leaf unfolding, and leaf coloration, depend primarily on the climatic conditions (Barichivich et al., 2013; Cleland, Chuine, Menzel, Mooney, & Schwartz, 2007; Richardson et al., 2012). On the other hand, plant phenology influences climate, defining the biosphere-atmosphere boundary conditions which influence surface albedo and terrestrial water cycle dynamics (Brovkin, 2002; Pielke et al., 1998), as well as the atmospheric chemistry through the exchange of several compounds (Guenther et al., 2006; Lathière et al., 2006; Messina et al., 2016; Petroff, Mailliat, Amielh, & Anselmet, 2008). Therefore, reliable phenological models are needed to correctly quantifying gas exchanges between terrestrial vegetation and atmosphere.

Among common air gasses, ozone (O3) plays a pivotal role in the Earth system: in fact, it affects climate with a direct radiative forcing of 0.2–0.6 W/m2 (Ainsworth, Yendrek, Sitch, Collins, & Emberson, 2012; Myhre et al., 2013; Shindell et al., 2009, 2012) and the ecosystems, causing a reduction of carbon assimilation by the vegetation (Wittig, Ainsworth, Naidu, Karnosky, & Long, 2009) that accelerates the rate of rise in CO2 concentrations with indirect implications for climate change (Sitch, Cox, Collins, & Huntingford, 2007). Several studies show that tropospheric O3 could cause reductions in crop yield and forest production ranging from 0% to 30% (Anav, Menut, Khvorostyanov, & Viovy, 2011; Fares et al., 2013; Ren et al., 2007; Sitch et al., 2007; Tang, Takigawa, Liu, Zhu, & Kobayashi, 2013; Wittig et al., 2009). In addition, high O3 concentrations affect vegetation by decreasing foliar chlorophyll content into the leaves and photosynthesis, leading to an alteration of carbon allocation in the different pools (Bytnerowicz, Omasa, & Paoletti, 2007; Karnosky, Skelly, Percy, & Chappelka, 2007; Wittig et al., 2009). Besides, O3 accelerates leaf senescence (Gielen, Löw, & Deckmyn, 2007), changes susceptibility to abiotic and biotic stress factors (Karnosky et al., 2002) and makes sluggish or impaired response of stomata to environmental stimuli (Hoshika et al., 2015).

To date, the European standard used to estimate the negative effect of O3 on vegetation is the AOT40 (Directive 2008/50/EC); however, in the recent years, a new metric has been proposed, the Phytotoxic Ozone Dose (POD). Unlike AOT40, which only depends on the O3 concentration in the air, POD accounts for environmental conditions that influence the O3 uptake through the stomata (Anav et al., 2016; Emberson, Ashmore, Cambridge, Tuovinen, & Simpson, 2000; Matyssek et al., 2007; Musselman, Lefohn, Massman, & Heath, 2006; Paoletti & Manning, 2007; Tuovinen, Emberson, Simpson, 2009). For risk assessment over Europe, the O3 uptake by plants is based on the multiplicative Jarvis’ algorithm for calculation of stomatal conductance (Jarvis, 1976) computed using the Deposition of Ozone for Stomatal Exchange (DO3SE) model (CLRTAP, 2015; Emberson, Ashmore, Simpson, Tuovinen, & Cambridge, 2001; Emberson et al., 2000). In addition to several meteorological variables, the DO3SE model requires a specific function describing the vegetation phenology which allows defining the temporal window for the O3 accumulation into the leaves (CLRTAP, 2015): specifically, it is used to compute the duration of growing season during which plants can uptake O3. To define the start (SOS) and end (EOS) of the growing season, the DO3SE model applies a simple latitude model: the SOS occurs at DOY (day of year) 105 at latitude 50°N and it is altered by 1.5 days per degree latitude earlier on moving southward and later on moving northward (CLRTAP, 2015). Similarly, the EOS is estimated as occurring at DOY 297 at latitude 50°N and is altered by 2 days per degree latitude earlier on moving northward and later on moving southward (CLRTAP, 2015). In addition, the effect of altitude on phenology is considered by assuming a later SOS and earlier EOS by 10 days for every 1,000 m above sea level (CLRTAP, 2015). As the latitude model uses a fixed time window for the O3 accumulation, it does not consider the interannual variability in the climate system as well as the influence of the climate change on the lengthening of the growing season and thus on the duration of the period during which plants can uptake O3. This limitation is strengthened by several recent studies which report earlier green-up dates and delayed dormancy dates because of changing climate (Jeong, Ho, Gim, & Brown, 2011; Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al.,2016; Menzel et al., 2006; Schwartz, Ahas, & Aasa, 2006; Wolkovich et al., 2012).

The past decade has seen a particularly rapid increase in the number of launched satellites, as well as an improvement in both spatial and temporal resolution of the data they produce. In this regard, remote sensing is a valuable tool for plant phenology estimation. In fact, the seasonal cycle of the normalized difference vegetation index (NDVI) (e.g. Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al., 2016; Myneni, Keeling, Tucker, Asrar, & Nemani, 1997; Piao, Fang, Zhou, Ciais, & Zhu, 2006; Shen et al., 2014; Wu & Liu, 2013; Zeng, Jia, & Epstein, 2011) as well as time evolution of the leaf area index (LAI) (e.g. Anav et al., 2013; Murray-Tortarolo et al., 2013; Piao, Friedlingstein, Ciais, Viovy, & Demarty, 2007; Zhu et al., 2016) has been widely used to estimate the timing of phenological events. In addition, because of some uncertainty associated with NDVI estimation (Huete et al., 2002; Jin & Eklundh, 2014; Mutanga & Skidmore, 2004), several vegetation indices (VIs) (e.g. Enhanced Vegetation Index, Green Chromatic Coordinate, Plant Phenology Index) have been developed and used to infer plant phenology (Churkina, Schimel, Braswell, & Xiao, 2005; Keenan et al., 2014; Shen et al., 2016); the comparison of green-up dates estimated by different vegetation indices, showed a good correspondence in the spatial pattern with a slight temporal difference (Keenan et al., 2014; Peng et al., 2017; Shen et al., 2016).

In this frame, this work aims to assess how much the stomatal O3 flux changes when different phenological models relying on satellite retrievals are used to estimate the start and the end of growing season. Here, using a regional climate model and a chemistry transport model, we compare for the year 2011 the stomatal O3 fluxes computed using different algorithms and methodologies to estimate the green-up and dormancy dates over a large region extending from Northern Africa to the Scandinavian region and part of the Russia.


2.1 Atmospheric regional climate model and chemistry transport model

To produce the meteorological variables needed to run the chemistry transport model and thus estimate the stomatal O3 flux, we used the Weather Research and Forecasting (WRF) model (v3.6); it is a limited-area, nonhydrostatic, terrain-following eta-coordinate mesoscale model (Skamarock & Klemp, 2008) widely used worldwide for climate studies. The study area (Figure 1) was selected in order to have a large latitudinal extension with a fine spatial resolution (16 km).

The initial and boundary meteorological conditions are provided by the European Centre for Medium-range Weather Forecast (ECMWF) analyses with a horizontal resolution of 0.7° every 6 hr (Dee et al., 2011).

The chemistry transport model used in this study is CHIMERE (v2013b), an Eulerian model developed to simulate gas-phase chemistry, aerosol formation, transport and deposition at regional scale (Menut et al., 2013). To accurately reproduce the gas-phase chemistry, emissions must be provided every hour for the specific species of the chemical mechanism. For studies over Europe, the EMEP inventory (Vestreng et al., 2009) is usually used for anthropogenic emissions of NOx, CO, SO2, PM2.5, and PM10, while biogenic emissions are calculated through the MEGAN model (Guenther et al., 2006). CHIMERE is widely used and validated (http://www.lmd.polytechnique.fr/chimere/CW-articles.php) and has already been used for O3 risk assessment on European vegetation; further details are described in Anav et al. (2016).

For both WRF and CHIMERE, we performed a simulation covering the whole year 2011, with a short spin up of 2 months to initialize all the fields.

2.2 Land cover

Land cover is a key variable in stomatal O3 flux assessment as each vegetation category has specific properties and ranges for the stomata opening defined by some climatic variables. The WRF preprocessing system provides land use and land cover data (24-category classification) from the U.S. Geological Survey (USGS) Global Land Cover Characterization (GLCC) dataset; with this preprocessing the land cover, derived from the Advanced Very High Resolution Radiometer (AVHRR) data with a resolution of 30 arc-seconds (<1 km) (Loveland et al., 2000; Sertel, Robock, & Ormeci, 2010), is regridded to any region of interest (Fig. S1a).

The output data represent the vegetation fraction (as percentage) of each of the 24 categories. We defined cropland the grid points where the cover of crops is >90% (Fig. S1b); these points are excluded by our analysis as they do not represent natural ecosystems and thus plant phenology is driven by human management. In the remaining points, we retrieve the dominant forest cover from the vegetation fraction. Finally, to convert the USGS categories into the DO3SE land cover (Figure 1), for which all the bio-climatic ranges are defined for the different land covers types, we use the Köppen-climate classification (Kottek, Grieser, Beck, Rudolf, & Rubel, 2006). We compute the climate zones over our domain (Fig. S2) from the temperature and precipitation simulated by WRF.

2.3 Modeling the stomatal ozone flux: the DO3SE model

The leaf-level stomatal conductance (gsto, in mmol O3 m−2 s−1) is estimated using several limiting functions which consider the variation in the maximum stomatal conductance (gmax) with photosynthetic photon flux density (flight), surface air temperature (ftemp), vapor pressure deficit (fVPD), and volumetric soil water content (fSWC). These functions integrate the effects of multiple climatic factors, vegetation characteristics and local features on the stomatal conductance; they vary between 0 and 1, with 1 meaning no limitation to stomatal conductance (e.g. Emberson et al., 2000).

In addition, the DO3SE model requires another function describing the phenology of vegetation (fphen); this function is used to compute the duration of growing season during which plants can uptake O3. As already described above, to define the start (SOS) and end (EOS) of the growing season, the DO3SE model applies a simple latitude model (CLRTAP, 2015). In this study, in addition to the latitude model, we use the bi-weekly composited GIMMS LAI3g (Zhu et al., 2013) and NDVI3g (Pinzon & Tucker, 2014) data to extract the date of start and end of vegetation growing season (discussed below).

Given the above-mentioned limiting functions, the stomatal conductance is computed as following:


where gmax is the maximum stomatal conductance of a plant species to O3, and fmin is the minimum stomatal conductance expressed as a fraction of gmax (Emberson et al., 2000).

Finally, we compute the POD using a threshold of 0 nmol m−2 s−1, because a large uncertainty still exists on the amount of O3 in the mesophyll that is detoxified without inducing any injury (De Marco et al., 2016). The POD0 (mmol O3 m2) is computed as following:


where Rb represents the quasi-laminar resistance and depends on wind speed and leaf dimensions, Rc is the canopy resistance (Simpson, Emberson, Ashmore, & Tuovinen, 2007), dt is 1 hr, and gsto is the stomatal conductance to ozone.

2.4 Satellite LAI and NDVI data

The LAI dataset used in this study (LAI3g) was generated using an artificial neural network from the third generation of GIMMS AVHRR NDVI data for the period July 1981 to December 2013 at 15-day frequency and with a 1/12 degree of spatial resolution (~8 km) (Zhu et al., 2013). Further details on the LAI3g and the comparison with other satellite products are provided in Zhu et al. (2013) and Fang et al. (2013).

Considering the year 2011, firstly we regridded the LAI3g data from its native grid to our domain using a bilinear interpolation. Then, the green-up and dormancy dates were calculated based on the LAI seasonal amplitude: in fact, LAI has been shown to have a normal distribution over the year in northern latitudes (Zhang, Anderson, Barlow, Tan, & Myneni, 2004), so we consider the start of the growing season to be 20% of the maximum amplitude (Anav et al., 2013; Murray-Tortarolo et al., 2013).

In the following analysis, we mask out grid points with a small changes in LAI over the year, i.e. those points where the difference between the maximum and minimum LAI amplitude is <0.5. In fact, these points represent crops, evergreen forests or mixed forest with a small deciduous component (Anav et al., 2013; Murray-Tortarolo et al., 2013) (Figure 1). In case of Mediterranean evergreen forests, we assign a growing season length of 365 days, while crops are excluded by this study.

For any grid point (x,y) of our domain covered by forests, we calculated a critical threshold value (CTx,y) above which we assume the plants to be photosynthetically active and thus able to uptake O3:


where and represent the minimum and maximum LAI over the year 2011 for the grid cell (x,y). From this threshold, the green-up date is then calculated as the day when LAI starts to increase, while the dormancy date is defined as the day when LAI down crossed the threshold (Anav et al., 2013; Murray-Tortarolo et al., 2013; Piao et al., 2007).

We also used the third generation of NDVI records (referred as NDVI3g) to estimate the plant phenology; this dataset has a spatial resolution of about 8 km with a bi-weekly time frequency covering the past three decades (Pinzon & Tucker, 2014).

Several methods have been developed to extract green-up and dormancy dates from NDVI data (Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al., 2016, White et al., 2009). Following Piao et al. (2006), for any grid point (x,y) we first calculated the NDVI ratio from a 20-year averaged NDVI time-series (1992–2011) to determine the average onset dates of vegetation green-up and dormancy:


where t is time (temporal resolution of 15 days) and NDVI(t) is the bi-weekly NDVI3g value at time t.

During this phase, original NDVI3g data are regridded to our model grid using a bilinear interpolation (Figure 1) and the pixels influenced by snow coverage are screened out and replaced by NDVI of the temporally nearest snow-free date, allowing to restrict the estimated SOS/EOS from going beyond thermal growing season (Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al., 2016). To identify the pixels influenced by snow during the 1992–2011 time period, we used daily air temperature from ERA-INTERIM dataset (Dee et al., 2011), previously downscaled to our domain via a dry-adiabatic lapse rate. After this preprocessing of the NDVI3g data, we used the four methods (i.e. POLYFIT, HANTS, DOUBLE-LOGISTIC and PIECEWISE logistic) described by Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al., 2016 to estimate the SOS and EOS dates. Detailed information on the data processing and the sampling of the NDVI ratio function to estimate the phenology can be found in Piao et al. (2006) and Liu, Fu, Zeng, et al., 2016, Liu, Fu, Zhu, et al., 2016.

2.5 Validation against measurements

In order to analyze the ability of CHIMERE to reproduce the spatial distribution of surface ozone concentration over Europe, we compare the model against station data. In situ measurements were obtained from the European air quality database (AirBase) and maintained by the European Environment Agency (EEA) (http://acm.eionet.europa.eu/databases/airbase/). Although the AirBase stations are classified by type (i.e. rural, suburban, urban), in the following validation we do not remove the urban stations in order to have a full picture of models skills, even if it can increase the model-data mismatch.

Considering the phenology, the model validation is performed in nine forest sites (Figure 1) belonging to the European fluxes network (http://www.europe-fluxdata.eu). The choice of these specific sites is due to the multiple requirements of having full year data coverage with a wide range of forest ecosystems across different latitudes, landscapes and climatic zones. Specifically, the sites refer to different plant types covering the most important European forest ecosystems (Table S1).

For the validation of the DO3SE model, we perform a correlation analysis between the mean seasonal cycle of simulated O3 fluxes and gross primary production (GPP) estimates. It should be noted that GPP is not directly measured by the eddy covariance technique but it is estimated using a partitioning technique (Reichstein et al., 2005). Nevertheless, as the exchange of gases between atmosphere and terrestrial vegetation is regulated by stomata opening, air pollutants may take advantage of the stomatal aperture to enter leaves, suggesting that the temporal evolution of O3 and CO2 uptake are consistent (Cieslik, Omasa, & Paoletti, 2009); for this reason, the correlation coefficient computed between O3 uptake and gross photosynthesis (i.e. CO2 flux) can be regarded as a good metric to assess the ability of the DO3SE model to reproduce reasonable temporal patterns of O3 fluxes (Fares et al., 2013; Proietti, Anav, De Marco, Sicard, & Vitale, 2016).


3.1 Performances of atmospheric chemistry model

Figure 2 shows the validation of daily surface O3 concentrations against AirBase measurements for the temporal periods April–May–June (AMJ) and July–August–September (JAS) 2011.

Compared to the in situ data, CHIMERE is able to reproduce the background concentrations in Europe in the two analyzed seasons. The high concentrations typical of the Mediterranean region and the low concentrations of Central Europe and Russia are fairly well reproduced for both AMJ and JAS. Similarly, in central Europe the decrease in O3 concentration between AMJ and JAS caused by the NOx titration is well captured. In contrast, as already discussed by Terrenoire et al. (2015) and Anav et al. (2016), the model exhibits higher values and a more pronounced spatial gradient than the observations. This is particularly clear in the Mediterranean coasts of Spain, over the whole Italy and the mountainous regions, where the model is between 5 and 15 ppb higher than the reference data and during JAS in England and Northern Europe. The overall AMJ bias (reference-model), computed comparing model's results with 1043 in situ sites, is −5.8 ppb, while the mean daily root mean standard error (RMSE) is 11.7 ppb and the correlation is 0.72. Similarly, during JAS, from the comparison with 1048 sites, we find a bias of −9.1 ppb, a RMSE of 13.2 ppb and a correlation of 0.63.

3.2 Differences in the spatial SOS and EOS patterns

All the algorithms used to estimate the SOS converge toward a similar latitude-dependent pattern, although we find some remarkable differences in the green-up dates among the different methods (Figure 3). In particular, except for mountainous regions, the methods show that the mean green-up date is progressively delayed with increasing latitude and increasing continentally. Because of low temperatures, we find the latest dates of green-up in northern Scandinavia and Russia, around the Barents Sea, and over the mountainous regions, while the mild temperatures typical of the Mediterranean region promote earlier green-up dates with respect to boreal and tundra regions.

The mean green-up date estimated by the latitude model is DOY 104 (14 April), while with all the other methods the vegetation turned green much earlier, up to 1 month (NDVI3g-HANTS DOY 76, NDVI3g-PIECEWISE DOY 77, NDVI3g-POLYFIT DOY 84, LAI3g DOY 98, NDVI3g-DOUBLE-LOGISTIC DOY 103) (Fig. S3).

In order to assess the spatial consistency between the different methods and the standard latitude model used by the DO3SE, we compute the spatial correlation (Fig. S4). In general, considering the latitude model as reference, there is a reasonable agreement between the different algorithms and the latitude model: we find the highest spatial agreement with the NDVI3g-DOUBLE-LOGISTIC (r = 0.79) and NDVI3g-PIECEWISE (r = 0.78), while the lower correlation occurs with the NDVI3g-HANTS method (r = 0.2).

The vegetation dormancy occurs in reverse order of the green-up onset, namely the green-up wave progresses northwards and the dormancy wave progresses southwards. In the high latitude regions, characterized by low temperature and snow cover, the dormancy starts earlier than in Southern Europe where mild temperatures promote a longer growing season. The different methods capture the progressively early dormancy with increasing latitude, although, compared to the other methods, the NDVI3g-HANTS algorithm shows a completely different picture with delayed offset dates in the Arctic region (Figure 4). Similarly, the NDVI3g-PIECEWISE remarkably differs from other estimates in all the continental Europe. The mean dormancy date computed using the latitude model is DOY 294 (21 October), while the other methods show both earlier (NDVI3g-HANTS DOY 279, NDVI3g-POLYFIT DOY 280, NDVI3g-DOUBLE-LOGISTIC DOY 284) and later offset dates (LAI3g DOY 297, NDVI3g-PIECEWISE DOY 313) (Fig. S3).

The spatial agreement between the latitude model and the EOS dates computed using the other methods is much poorer than the SOS estimates (Fig. S4): similarly to SOS, we find the highest spatial correlation with NDVI3g-PIECEWISE (r = 0.47) and NDVI3g-DOUBLE-LOGISTIC (r = 0.44), while because of the large disagreement in the high latitude snow-dominated regions the correlation between the latitude model and NDVI3g-HANTS is negative (r = −0.16) (Fig. S4).

The effect of different estimates in the SOS and EOS dates leads to a remarkable difference in the growing season length (GSL) and thus in the period during which plants are susceptible to be damaged by O3. The mean GSL estimated using the latitude model is of 190 days, while the other algorithms show both shorter (NDVI3g-DOUBLE-LOGISTIC 181 days) and longer (NDVI3g-POLYFIT 196 days, LAI3g 199 days, NDVI3g-HANTS 203 days, NDVI3g-PIECEWISE 236 days) growing seasons (not shown).

3.3 Effect of phenology on stomatal O3 flux

Interestingly, despite the green-up and dormancy dates show remarkable spatial differences between the various methods, the stomatal O3 flux has a similar spatial pattern, although the magnitude of the fluxes slightly differs (Figure 5). This is confirmed by the relative high spatial correlation coefficients computed using the latitude model as reference: in fact, the correlation ranges between 0.68 in case of LAI3g and 0.75 for NDVI3g-POLYFIT, NDVI3g-HANTS and NDVI3g-DOUBLE-LOGISTIC methods (not shown).

Besides, it is noteworthy that the maximum change in the domain-averaged magnitude of fluxes computed comparing the various approaches is 25%: we find the minimum flux for the latitude model (7.455 TgO3/year), while with the other algorithms the POD0 values are very close each other ranging between 8.056 TgO3/year for NDVI3g-DOUBLE-LOGISTIC and 9.343 TgO3/year in case of NDVI3g-PIECEWISE. If we exclude the latitude model from the analysis, this result is consistent with the GSL, namely the shorter is the length of growing season the smaller is the cumulated stomatal O3 flux.

Despite the latitude model has, on average, a longer growing season than the NDVI3g-DOUBLE-LOGISTIC (9 days), its POD0 is smaller (0.6 TgO3/year): this pattern derives from the sum of points with smaller and larger POD0 values. In particular, in the Iberian Peninsula, Northern of the Alps and in the Southern part of the Sweden the NDVI3g-DOUBLE-LOGISTIC model shows much earlier green-up dates than the latitude model and thus plants uptake a much larger amount of O3 which, in turn, affects the domain-averaged POD0.


Plant phenology is a key variable in the atmospheric climate and chemistry models, as it regulates many processes in the planetary boundary layer. Using two satellite datasets and six widely used methods to extract the start and the end of the growing season, our results indicate a common spatial pattern in the cumulated stomatal uptake of O3, although SOS and EOS estimates remarkably differ in terms of average DOY. This finding suggests that, on regional scale, climate is the main driver of the uptake of gases from the atmosphere to the vegetation (Ahlström et al., 2015; Nemani et al., 2003; Seddon, Macias-Fauria, Long, Benz, & Willis, 2016). This result is consistent with Anav et al. (2015) which showed how the gross uptake of CO2 by terrestrial ecosystems is quite homogeneous among global models that use different parameterizations for the photosynthesis. In other words, despite models make use of many various numerical schemes and have different climate sensitivities (Piao, Ciais, & Friedlingstein, 2013), they all converge to a common spatial pattern of gross primary production (Anav et al., 2015). In addition, even a bias in the climate models does not affect both the phenology and the seasonal carbon fluxes, but it significantly controls the magnitude of carbon assimilation by vegetation (Dalmonech et al., 2015).

It should be noted that the above comparison in SOS/EOS estimates does not establish the correctness of any one method, but it only provides a relative difference in the timing of events, with a direct impact on the cumulated stomatal O3 flux. Therefore, such an intercomparison would have no rational basis for selecting one method over another method. In the integrated assessment modeling (IAM), the latitude model is widely used to retrieve the plant phenology and thus estimate the impacts of O3 on vegetation (CLRTAP, 2015). This method, in fact, may work on large regions or when no satellite retrievals or other data are available, e.g. in case of prognostic simulations (i.e. scenarios). However, we believe that a comparison with other estimates is needed to assess how much the latitude model, and hence the stomatal uptake, differs when other phenological models are used.

To assess the performances of the different phenological models in reproducing the observed green-up and dormancy dates, in Figure 6 we compare the SOS and EOS estimates with GPP measurements collected in four in situ sites having a strong GPP seasonality. In Hyytiala, a stand of Pinus sylvestris in Finland (Kramer et al., 2002; Tanja et al., 2003), the GPP starts to increase at DOY 106. The NDVI3g-HANTS (DOY 69) and the NDVI3g-PIECEWISE (DOY 84) predict a too early SOS, while the LAI3g (DOY 120), the latitude model (DOY 125), and the NDVI3g-DOUBLE-LOGISTIC (DOY 128) well capture the SOS; in contrast, the NDVI3g-POLIFIT (DOY 155) estimates a too late SOS. Considering the EOS, the NDVI3g-POLIFIT (DOY 282), LAI3g (DOY 300), and NDVI3g-HANTS (DOY 316) are reasonably scattered around the EOS estimated through the GPP data (DOY 294), while the NDVI3g-DOUBLE-LOGISTIC (DOY 268) and the latitude model (DOY 271) predict a too early EOS. Similarly, in Sodankyla, a Pinus sylvestris forest site in arctic region of Finland (Kramer et al., 2002; Tanja et al., 2003), the NDVI3g

EVI3702/101/3/2017 3 1. INTRODUCTION AND WELCOME Welcome to the module Law of Evidence: the presentation and assessment of evidence! You will find a similar word of welcome in the introduction to your study guide. Please read both these introductory words of welcome with care, and note that, should the information they impart differ, you should follow the information in this tutorial letter. The Law of Evidence is presented in two modules, namely Evidence: admissibility of evidence(EVI3701) and Evidence: the presentation and assessment of evidence(EVI3702). This tutorial letter pertains to EVI3702 only, for both semesters of 2017. Both modules are normally taken during the third year of study. We are pleased to welcome you to this module, and hope that you will find it both interesting and rewarding. We will do our best to help you achieve success in your study of this module. You will be well on your way to success if you start studying early in the semester/year and complete the assignments with due diligence. You will receive a number of tutorial letters during the year. A tutorial letter is our way of communicating with you about teaching, learning and assessment. Tutorial letter 101 contains important information about the scheme of work, resources and assignments for this module. We urge you to read this tutorial letter carefully and to keep it at hand when working through the tutorial material, preparing the assignments, preparing for the examination and addressing questions to your lecturers.

Leave a Comment


Your email address will not be published. Required fields are marked *