STICS Soil-crop model

Recent advances in intercropping modelling: the new version of the STICS soil-crop model simulates consistently a wide range of bi-specific annual intercrops. Eric Justes, Remi Vezy, Sebastian Munz, Kirsten Paff, Laurent Bedoussac, Noémie Gaudio, Patrice Lecharpentier, Dominique Ripoche, Marie Launay

By using the crop model STICS in the present study, we aimed to provide a better understanding of how crop manage- ment can influence nitrate leaching and, in consequence, affect the quality of ground water. Particularly, the main question was the following: was the change of crop from potato to sugar beet, and of the origin of irrigation water able to explain the decrease of nitrate concentration observed in the ground water? Because there are interactions between crop rotation, crop management and pedoclimatic conditions, it is not always easy to distinguish the individual roles of each of these components and their interactions on nitrate leaching only based on field measurements. The use of a soil –crop model could be very useful to diagnose the impacts of crop type, N-fertilization, irrigation, and climate on nitrate con- centrations of drained water by making complementary simulations with combinations of agricultural practices and climatic conditions not occuring during the field experiment. In the field studies in the Vitoria-Gasteiz region, potatoes (Solanum tuberosum L.) were grown in 1993 when the nitrate concentration in ground water was maximum and sugar beet (Beta vulgaris L.) in 2002 when the nitrate concentration had decreased to about 60 mg NO − 3 L − 1 . The aims of our study were:

Robust simulation responses for economic modelling require observations for estimation or validation of how agents react to changes in market and policy signals. Most often, at least for large-scale analysis, solely time series data for larger administrative units provide the neces- sary variance in signals and responses. For Pan-EU economic analysis, already a sub-national regional resolution as in CAPRI (Common Agric. Policy Regional Impact model, Britz et al., 2007) is, therefore, rather unique. With higher spatial variance inside administrative units, estimates of environmental impacts based on regional averages may be considerably biased under non-linear dependencies between these impacts and soil parameters, climate or farming practice. Certain impacts can only be assessed in their proper spatial setting such as the relation between nitrate leaching and drinking water resources or the effects of land use on specific habitats. Statistical down-scaling provides, therefore, a bridge between large-scale economic and environmental analysis in agriculture. The behavioral response regarding crop shares, yields or animal stocking densities is simulated for administrative units with the economic model. The downscaling tool then consistently distributes these changes to geo- referenced units below the administrative level. Environmental impacts are then analysed with indicator calculators or simulated with biophysical models at an appropriate spatial resolution. A work package of the project ‘CAPRI-Dynaspat’ contributed a spatial down-scaling tool to CAPRI. It established the necessary geo-referenced data bases and developed methodologies and software to consistently dis-aggregate for the whole EU27 all major results from the CAPRI modelling system to about 150.000 clusters of 1×1 km grid cells (Leip et al., 2007). Methods

b Institut National de Recherche Agronomique (INRA), Unité Climat, Sol et Environnement, Domaine St Paul, Site Agroparc, 84914 Avignon Cedex 9, France c INAPG, Chaire de Bioclimatologie, Département AGER, 16 rue Claude Bernard, 75231 Paris Cedex 05, France (Received 16 June 2003; accepted 14 May 2004) Abstract – Agriculture is still accounted for in a very simplistic way in the land-surface models which are coupled to climate models, while the area it occupies will significantly increase in the next century according to future scenarios. In order to improve the representation of croplands in a Dynamic Global Vegetation Model named ORCHIDEE (which can be coupled to the IPSL 1 climate model), we have (1) developed a procedure which assimilates some of the variables simulated by a detailed crop model, STICS, and (2) modified some parameterisations to avoid inconsistencies between assimilated and computed variables in ORCHIDEE. Site simulations show that the seasonality of the cropland- atmosphere fluxes of water, energy and CO 2 is strongly modified when more realistic crop parameterisations are introduced in ORCHIDEE. A more realistic representation of wheat and corn croplands over Western Europe leads to a drying out of the atmosphere at the end of summer and during autumn, while the soils remain wetter, specially at the time when winter crops are sowed. The seasonality of net CO 2 uptake fluxes is also enhanced and shortened.

There was no significant difference in SMN values at harvest or in nitrate leaching for the different main crops in the study area, although large between-field variations were observed. Nitrogen management in this part of the alluvial floodplain was not effec- tive and hence the nitrate concentrations in drainage water under crops were too high. Drainage and nitrate concentration values var- ied widely from one field to the next, depending on the previous crop, agricultural practices (with or without irrigation) and annual climate conditions. For some fields, the average annual nitrate con- centration in drainage water was greater than 200 mg NO 3 − L −1 and nitrate leaching exceeded 100 kg N ha −1 . Analyses of temporal and spatial variability in nitrate leaching showed that the pattern of nitrate leaching was extremely specific and irregular (spatially and temporally) and also that the SMN content at the end of autumn, before the winter drainage period, was the most significant factor explaining this variability. For the study area, this means that N management must be aimed at reducing SMN as much as possible in November. This means that N fertilization for the next main crop must be adjusted by taking into account the residual SMN at the beginning of the crop season (soil analysis may be necessary) and by planting catch crops to decrease SMN before the winter.

Because current farming systems have many issues, particularly related to their negative impacts on the environment, it is necessary to operate a transition toward more sustainable agriculture and redesigning cropping systems. Crop diversification, which acts on planned biodiversity as well as associated biodiversity, is a solution for designing disruptive sustainable farming systems, able to provide several ecosystem services. Designing such farming systems including various levels of crop diversification via crop rotation and intercropping needs crop modelling because it allows assessing ecosystem services provided by a wide range of virtual systems, on different climate-soil contexts, on a long-term basis. Consequently, the objective of this study is to define detailed, by the conceptualisation of an “ideal” system, characteristics that a crop model should have to simulate crop diversification and its consequences on various ecosystem services relevant in Brittany region. For this, among 3 crop models (APSIM, DSSAT, STICS), we select the best crop model according to the detailed characteristics proposed in "ideal" system and we test its relevance in Brittany region. APSIM comes out as the crop model best suited to respond to our objective and to simulate and follows indicators related to the provision of ecosystem services. However, despite encouraging results, APSIM shows some limits to getting use to Britany pedoclimate context principally because a lack of adapted cultivars and complete information about initial conditions of soil. Thus, given this weakness, it is necessary to set up an adaptation step of the crop model and to complete this work with other simulations whose outputs will be compared with more observed data. In this way, it could be established whether APSIM is a good tool for designing systems based on crop diversification in the Brittany Region.

Materials and Methods The data set used for modelling comprised of four years of wheat (Triticum turgidum L.) and pea (Pisum sativum L.) field data from Auzeville, France with multiple levels of nitrogen fertilizer, and four years of barley (Hordeum vulgare L.) and pea field data from Angers, France (Corre-Hellou, 2005), which in some years included two levels of nitrogen fertilizer and two different plant densities of the intercrops. The sole crop trials were used for calibration and the intercrop trials for evaluation, except for a subset of intercrop data that was used to calibrate the parameters unique to the intercrop model. The assumption was that parameters common to both sole and intercropping, such as plant-soil interactions and phenology, would be the same for both. The optimization method used for calibration was based on Wallach et al. (2011). The parameters were broken down into 15 groups (16 for pea to include nitrogen fixation) for calibration, each corresponding to a different process.

Vezy R et al. (2020) Implementation of new formalisms in STICS for intercropping modeling, iCROPM, Montpellier, France. Wallach D et al. (2011) A package of parameter estimation methods and implementation for the STICS crop-soil model. Environmental Modelling & Software, 26(4): 386-394. Figure 1: Comparison of simulated observed grain yield for sole crop winter wheat, winter pea, spring barley, and spring pea.

Figure 1. Simulated versus observed yields. a. Individual fields results, with identification of additional limiting factors. b. Average yield for each class of fields (358 fields). Legend: St sterility, W weeds, S salinity, sD small measured plant density, C calibration, V validation, aV additional varieties (similar to known varieties), H high measured harvest index, O high measured soil organic nitrogen, T high measured tiller

b CIRAD, PERSYT Department, , 34980 Montpellier, France Keywords: STICS model, simulation study, cover crop management, next cash crop, water, nitrogen Introduction Cover crops provide multiple ecosystem services such as reducing nitrate leaching, producing “green manure” effect, improving soil physical properties, increasing carbon storage in the soil and controlling pests, diseases and weeds (Justes et al., 2017). Cover crops increase evapotranspiration by increasing cover transpiration and decrease soil evaporation, and then they reduce water drainage in temperate climates (Meyer et al., 2019). However, the equilibrium of these processes and ecosystem services provided depends on cover crop management, climate and soil type. No consensus exists on the impact of cover crops on soil water availability for the next cash crop. Dynamic soil-crop models can be a powerful tool to estimate water fluxes that are difficult to measure in field experiments, such as drainage, evaporation, and transpiration. They can also be used over long climatic series for evaluating their variability versus weather and a wide range of management practices (Bergez et al. 2010). We hypothesis that the cover crop management must take into account soil and climate context to maximize the multiple ecosystem services and in the same time reduce the negative impact of cover crops on soil water balance and on the next cash crop. Our goal was to analyse by simulation the best cover crop managements according to soils and climates in the Adour-Garonne catchment.

GreenLab is such type of model. It takes its origin in the AMAP architectural models (de Reffye, 1988) but its ecophysiological concepts are inspired from those classically used in process-based models (Monteith 1977; De Wit 1978; Howell and Musick 1985; Marcelis et al 1998 or Qi et al 2005 for sugar beet), except that the dynamics of source-sink interaction is described at the level of organs according to their rhythm of appearance, see de Reffye and Hu (2003). The model does not claim to be fully mechanistic with regards to physiological and biophysical processes and fluxes involved in plant growth but a particular care is taken to follow empirically the dynamics of the carbohydrate budget, production and allocation, see (Yan et al 2004).

on  earthworm  communi8es  and  soil  physico-­‐chemical  proper8es  in  silt  loam  arable  soil  (Belgium)   A.  Lem8ri,  G.  Colinet,  T.  Alabi,  B.  Bodson,  D.  Cluzeau,  Y.  Brostaux,  C.  Olivier,  J.  Pierreux,  E.  Haubruge  &  F.  Francis   University  of  Liege  –  Gembloux  Agro-­‐Bio  Tech   Soil  and  Water  Systems  Unit  -­‐  Func8onal  &  evolu8onary  Entomology  Unit  

Creep of model piles in frozen soil Parameswaran, V. R. https://publications-cnrc.canada.ca/fra/droits L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

studied by treating the ensemble as the full population of models and drawing subsamples from that population. The conclusions have been that prediction error decreases systematically as the number of models increases. Li et al. (2015) suggested that eight models would be suffi- cient to obtain errors of e ‐mean below 10% of observed yield. All of these studies have been empirical, based on a single MME study. The general behavior of crop ensemble predictors has not been addressed. Studies in other fields, including group intelligence (Surowiecki, 2005), hydrologic modeling (Duan, Ajami, Gao, & Sorooshian, 2007), air quality modeling (Solazzo & Galmarini, 2015), and climate modeling (Tebaldi & Knutti, 2007), have also found that averaging over multiple opinions or solutions can give good predictions, often better than any individual model. The basis for using MME predictors has received particular attention in the field of climate modeling (Hagedorn, Doblas ‐Reyes, & Palmer, 2005; Weigel, Liniger, & Appenzeller, 2008). However, the con- text there is quite different than for crop models; for example, in climate modeling each MME member is often itself an ensemble based on a sin- gle model with different initial conditions (DelSole, Nattala, & Tippett, 2014) whereas in crop modeling, each model normally provides a single simulation, a major interest in climate modeling is in probabilistic predic- tions rather than the deterministic predictions of crop models (DelSole et al., 2013; Wang et al., 2009) and in climate modeling spatial patterns of prediction play an important role (DelSole et al., 2013).

Daily meteorological inputs from automatic measurements (Temperature, Soil moisture content) in each soil layer
Residue input frequency: once a cultural year (at harvest) Fig. 2: Inputs (in green) and layers of the soil heterotrophic respiration model

• Yields and quality of the harvested products • Presence of diseases and pests • Nitrogen uptake by the plants Soil-water-plant continuum (after Zhuang et al., 2014) Green T.R., Ahuja L.R. & Benjamin J.G., 2003. Advances and challenges in predicting agricultural management effects on soil hydraulic properties.

3.3. Assessing the effects of crop residue management in the spatio­temporal dynamic context Several authors have studied the influence of crop residue management on soil hydraulic properties. Crop residue management, however, is not the only variable that differs among studies. First, temporal dynamics can overshadow punctual differences among treatments. Alletto et al. (2009) studied the parameters influencing the spatial and temporal variability of soil bulk density and hydraulic conductivity near saturation in France. They pointed out that the time of sampling was the first factor of variability for bulk density and for hydraulic conductivity function, whereas crop residue management was identified as the second or fourth factor. Some authors have reported a two-stage effect of tillage practices. Just after tillage, the soil surface structure of tilled systems is improved due to large open pores created by tillage (Messing et al., 1993). These newly formed pores, however, are relatively unstable, especially with conventional tillage. Under the influence of wetting/drying cycles and gravity, tilled soils tend to exhibit a decline in the number of macropores and their connectivity at the soil surface. This reduced macropore connectivity ultimately leads to a decrease in saturated hydraulic conductivity (Green et al., 2003). Alletto et al. (2009) observed that the most important increase in bulk density occurred between the two first measurements campaigns (between 10 and 51  days of measurement). Saturated hydraulic conductivity decreased by a factor 10 during the first month under conventional tillage at a depth of 15  cm. After this initial decrease, saturated hydraulic conductivity increased under both tillage practices over the cropping season, but the rate of increase was slower for conservation tillage. These seasonal changes also affect the water retention curve. While studying the temporal variation in soil physical properties in a spring ploughing in France aimed at managing maize residues, Alletto et al. (2015) also observed a decrease in the saturated water content during the growing season under conventional maize monoculture.

Figure 1. The photoperiod response factor in response to daylength using a photoperiod sensitivity parameter (PPSEN) ranging from 1 (blue) to 19 (red). 2.6. Growth Though there were no radiation use efficiency (RUE) data available for tef, it was assumed that, as a C4 crop, tef would have a higher RUE than wheat. Models for the C4 crops maize, millet, and sorghum all have higher aboveground RUE values than wheat models [20,44–46]. In order to calculate the DSSAT-Tef aboveground RUE, the percent difference between the DSSAT-CERES- Wheat and DSSAT-Sorghum aboveground RUE was used to estimate the ratio RUE ratio between wheat and a C4 crop. The percent change in RUE between the DSSAT-CERES-Wheat and DSSAT- Sorghum model was 18.5%. This percent increase was applied to the RUE of DSSAT-NWheat to estimate the RUE for tef. The RUE for DSSAT-NWheat was 3.2 g plant dry matter/MJ PAR, so the RUE value for DSSAT-Tef was set to 4.5 g/MJ. The NWheat model used a weighted mean temperature with emphasis on day time temperature [47]. To account for a higher base temperature for photosynthesis in tef, the threshold minimum temperature for photosynthesis was set to 4 degrees below the base temperature for phenology at 3.8 °C, the same difference, but for a different base temperature used in NWheat. These new parameter values were used in all tef simulations without any additional calibration.

Soil physical fertility: thesis project for water-soil-plant model improvement E. Beckers 1 * and A. Degré 1 1 Hydrology and Agricultural Eng., Environmental Science and Technology Department, Gembloux Agricultural University, 2 Passage des Déportés, 5030 Gembloux, Belgium (mail:beckers.e@fsagx.ac.be)

3 ITAP, Univ. Montpellier, INRAE, Institut Agro, Montpellier, France Abstract We consider a simple crop irrigation model and study the optimal control which consists of maximizing the biomass production at harvesting time. A specificity of this work is to impose a quota on the water used for irrigation, in a context of limited resources. The model is written as a 2d non-autonomous dynamical system with a state constraint, and a non-smooth right member given by threshold-based soil and crop water stress functions. We show that when the water quota is below the threshold giving the largest possible production, the optimal strategy consists of irrigating once. We then show that the optimal solution can have one or several singular arcs, and therefore be better than simple bang-bang controls, as commonly used. The gains over the best bang-bang controls are illustrated on numerical simulations. These new feedback controls that we obtain are a promising first step towards the concrete application of control theory to the problem of optimal irrigation scheduling under water scarcity.