*^M Agriculture Canada Research Direction generale Branch de la recherche Technical Bulletin 1992-2E Characterizing the water soil regime of the Canadian prairies Canada Digitized by the Internet Archive in 2012 with funding from Agriculture and Agri-Food Canada - Agriculture et Agroalimentaire Canada http://archive.org/details/characterizingso922dejo Characterizing the soil water regime of the Canadian prairies R. DE JONG, A. BOOTSMA, J. DUMANSKI, and K, SAMUEL Centre for Land and Biological Resources Research Ottawa, Ontario Technical Bulletin 1992-2E CLBRR Contribution No. 91-130 Research Branch Agriculture Canada 1992 Copiesofthispublicationareavailablefrom Director CentreforLandandBiologicalResourcesResearch ResearchBranch,AgricultureCanada RoomB149,K.W.NeatbyBuilding Ottawa,Ontario K1A0C6 ProducedbyResearchProgramService eMinisterofSupplyandServicesCanada1992 Cat.No.A54-8/1992-2E ISBN0-662-19398-9 Coverillustration Thedotsonthemaprepresent AgricultureCanadaresearchestablishments. INTRODUCTION La region agricole des Prairies canadiennes se caracterise par un climat continental au regime hydrique semi-aride a subhumide. Le gros de la production agricole de cette region se fait en culture seche. La production depend de la fiabilite des precipitations et de la capacite du sol d'emmagasiner de l'eau. Les pluies en saison de croissance sont souvent tres variables et il arrive souvent que cette variabilite augmente avec la diminution des precipitations totales. La capacite du sol d'emmagasiner les precipitations qui surviennent en dehors de la saison de croissance (hivernage) et la possibilite pour les cultures de disposer de cette eau sera de nature a reduire le risque que la production peut courir. II est prouve (Staple et Lehane, 1954; Lehane et Staple, 1965; Robertson, 1974; Bole et Pitman, 1980) que l'eau du sol emmagasinee au printemps et que les pluies tombees en saison de croissance sont pour la production cerealiere 1'equivalent de «precipitations efficaces» ou de source d'eau disponible. Les donnees sur l'apport, la distribution et la fiabilite de l'eau disponible pendant la saison de croissance et certains stades phenologiques cles du developpement des cultures sont done determinates pour assurer une agriculture viable et durable. Ces donnees sont essentielles aux agriculteurs pour les aider a planifier leurs techniques de gestion de l'eau du sol et aux decideurs pour leur permettre d'elaborer des programmes de securite du revenu et differents autres programmes. Les principaux sols agricoles de la region des Prairies sont les sols chernozemiques bruns, brun fonce, noirs et gris fonce et, dans une moindre mesure, quelques Luvisols gris (voir fig. 1). Ces sols sont repartis de fagon un peu concentrique autour de la zone la plus seche de la region, soit le sud-ouest de la Saskatchewan et le sud-est de 1'Alberta. Cette repartition temoigne d'une succession climatique liee a 1*augmentation de la disponibilite de l'eau qui rayonne vers l'exterieur de cette zone (Acton et al., 1980). Mais les sols de chacune de ces principales zones de sols peuvent etre de texture variable et done posseder des capacites de retention d'eau differentes. La quantite d'eau emmagasinee au printemps depend de la capacite de retention d'eau communement appelee capacite de retention d'eau disponible (CRED), des conditions meteorologiques au cours de la saison de croissance precedente, de l'hiver et des premiers mois du printemps, ainsi que de la pratique de la jachere l'annee precedente. La culture du ble de printemps, de l'orge, des oleagineux, des fourrages et la jachere sont les principales utilisations de sol de la region, mais leur repartition proportionnelle varie considerablement entre les diverses zones de sols. Le ble et la jachere (de meme que les parcours sur les sols marginaux) constituent les utilisations les plus frequentes dans les zones les plus seches et les plus chaudes, mais diminuent progressivement en allant vers le nord. L'orge, les fourrages et l'avoine montrent une repartition inverse caracterisee par une faible proportion dans le sud et des superficies generalement plus grandes en allant vers le nord. Les oleagineux montrent une repartition egale, mais moins reguliere. Les paturages ameliores sont generalement comparables dans toutes les regions, sauf dans le «grand» nord sur Luvisols gris ou ils occupent une superficie beaucoup plus grande de terres agricoles. ii Le rendement economique de 1*agriculture s'ecarte de ce modele et a tendance a temoigner plus fidelement des debouches et des possibilites du marche. Le capital investi et les ventes totales par hectare cultive sont plus eleves dans les zones de sols chernozemiques noirs et plus faibles vers le sud et le nord. Mais les depenses d' exploitation a l'hectare sont beaucoup moindres dans le sud de sorte que les marges beneficiaires brutes sont relativement comparables dans toutes les regions (sauf dans les zones de Luvisols gris fonce et gris ou elles sont plus faibles) (Huffman, 1988). Ces resultats temoignent de la tendance des agriculteurs de la zone des sols chernozemiques bruns qui est plus seche a utiliser de plus faibles niveaux d' intrants parce que l'apport et la repartition des precipitations au cours de la saison de croissance sont aleatoires, tout comme la reaction des cultures a l'apport d'intrants. Des etudes (Campbell et al., 1987, 1988; De Jong et Halstead, 1986 et Henry et al., 1986) ont revele que grace aux techniques actuelles de gestion, l'efficacite de l'eau conservee est beaucoup plus elevee que par le passe meme si la quantite est la meme. C'est a cause de la disponibilite de nouvelles varietes de culture et de la grande amelioration de la conduite des cultures. II est prouve (J. L. Henry, communication personnelle) qu'il faut environ 65 mm d'eau disponible dans la zone des sols bruns avant de pouvoir esperer obtenir un quelconque rendement du ble. Mais cette quantite diminue progressivement a 48, 41 et 38 mm dans les zones de sols brun fonce, noirs et gris fonce respectivement. Au-dela de ces seuils, on peut s'attendre a des augmentations de rendement, allant d' environ 9,2 kg/ha dans la zone des sols bruns a 12,5 kg/ha dans celle des sols gris, pour chaque millimetre supplementaire d'eau dont la plante en croissance peut disposer. Par consequent, les donnees sur les reserves d'eau du sol au printemps et sur la quantite de pluie prevue dans une region donnee sont tres importantes pour evaluer le rendement, et les donnees sur la quantite probable d'eau disponible dans le sol au printemps et de pluie en saison de croissance est essentielle a 1' evaluation du risque que court la production. Habituellement, on exprime l'apport d'eau par precipitation sous forme de moyennes arithmetiques a long terme meme si la pluie est un phenomene aleatoire ou stochastique. Cette etude s'ecarte d'une telle demarche en presentant 1 information en termes de ' niveaux variables de probabilite plutot que de moyennes (le seuil de probabilite de 50 % equivaut a la moyenne arithmetique) Elle compile egalement 1' information au moyen d'un . modele d'eau du sol qui integre la precipitation quotidienne pendant 30 ans a la capacite de retention d'eau du sol pour evaluer la quantite d'eau dont dispose chaque jour le ble de printemps a chaque periode de croissance. En outre, elle fournit des evaluations pour les divers stades phenologiques de croissance et diverses rotations de cultures. Ces evaluations portent sur differentes capacites de retention d'eau du sol dans chacune des zones de sols (zones de ressources agro-ecologiques) Les resultats . figurent sous forme de cartes montrant la repartition de l'eau du sol disponible a divers seuils de probabilite, et ce, pour les CRED les plus courantes de chaque region. Cette information permet de mieux comprendre la variabilite de l'eau disponible dans le sol dans les diverses regions des Prairies canadiennes et la probabilite d'un niveau donne de precipitation en saison de croissance. Elle temoigne de divers niveaux de risque que court la production et peut servir d'auxiliaire a la prise de decisions sur les pratiques culturales comme le semis, la jachere ou la rotation prolongee. Elle peut egalement servir a la prise de decisions en matiere de politiques et de programmes. 1 INTRODUCTION The agricultural region of the Canadian prairies has a continental climate, with semi-arid to subhumid moisture regimes. Most of the agricultural production in this region is based on dry- land farming. The risk of crop production depends on the reliability of precipitation and the ability of the soil to store water. Growing season rainfall is often highly variable and it is not unusual that rainfall variability increases as total precipitation decreases. The ability of the soil to store non- growing season (overwinter) precipitation and make it available to the growing crop will buffer the risk. It has been shown (Staple and Lehane, 1954; Lehane and Staple, 1965; Robertson, 1974; Bole and Pitman, 1980) that spring stored soil water and growing season rainfall can be considered together as available water or "effective precipitation" with regard to grain production. Therefore, information on the supply, distribution and reliability of available water during the growing season and during certain critical phenological stages in the development of the crop is critical for a viable and sustainable agriculture. The information is very important to farmers for planning soil water storage management techniques, and for policy makers for developing safety nets and other programs. The major agricultural soils in the prairie region are the Brown, Dark Brown, Black and Dark Gray Chernozemic soils, and to a lesser extent some of the Gray Luvisol soils (see Fig. 1) . These soils are distributed in a somewhat concentric manner around the driest area of the region, which is southwestern Saskatchewan and southeastern Alberta. This reflects a climate sequence related to increasing available water, which radiates outward from this area (Acton et al., 1980). Individual soils within each of these major soil zones however, can be of varying textures, and consequently varying soil water holding capacities. The amount of water stored in the spring depends upon the water holding capacity, commonly called Available Water-holding Capacity (AWC) the weather , conditions during the previous growing season, winter and early spring months and whether or not the land was fallowed the previous year. Spring wheat, barley, oilseeds, forages and summerfallow are the dominant land uses in the region, but' the proportional distribution of these varies considerably among the soil zones. Wheat and summerfallow (as well as rangelands on the marginal soils) are most common land uses in the driest and warmest areas, but decrease gradually to the north. Barley, forages and oats show a converse distribution, with low proportions in the south and generally increasing areas towards the north. Oilseeds are similarly distributed, but in a less regular pattern. Improved pasture is generally similar in all areas, except in the 'far' north on Gray Luvisol soils, where it occupies a much higher proportion of agricultural land. Economic performance of agriculture deviates from this pattern, tending to reflect market potentials and opportunities more closely. Capital investment and total sales per cultivated hectare are highest in the Black Chernozemic areas, and lower to the south and north. Operating expenses per hectare, however, are considerably lower in the south, with the result that gross margins are relatively similar in all regions (except for the Dark Gray and Gray Luvisol areas, where they are lower) (Huffman, 1988) This reflects the tendency of farmers in the drier, Brown Chern. ozemic soils to use lower levels of inputs because the supply and distribution of precipitation during the growing season is unreliable, as is crop response to inputs. Studies (Campbell et al., 1987; 1988, De Jong and Halstead, 1986 and Henry et al., 1986) have shown that with current management technigues, the efficiency of water conserved is much higher than that in the past even though the amount of water conserved is similar. This is due to the availability of new varieties and greatly improved crop management. It has been demonstrated (J. L. Henry, personal communication) that about 65 mm of available water is necessary in the Brown soil zone before one can expect any yield of wheat. This, however, decreases progressively to 48 mm in the Dark Brown, 41 mm in the Black and 38 mm in the Dark Gray soil zone. Beyond these thresholds, yield increases ranging from about 9.2 kg/ha in the Brown to 12.5 kg/ha in the Gray soil zone can be expected for each additional mm of water available to the growing plant. Thus, information on spring soil water storage and on how much rainfall can be expected in any given area is very important for estimating yield, and information on the probability of spring available soil water and growing season rainfall is fundamental to estimating production risk. Traditionally the supply of water through precipitation has been reported as long term arithmetic means, even though rainfall is a stochastic event. This study deviates from such an approach by presenting information in terms of varying levels of probabilities rather than means (the 50 per cent probability level is equivalent to the arithmetic mean) Also, the information is compiled using a soil water model. which integrates daily precipitation for a 30 year period, with the water storage capacity of the soil, to estimate the amount of water available to spring wheat for each day of each growing period. Furthermore, estimates are provided for different phenological growth stages, and for different crop rotations. These estimates are made for different soil water-holding capacities, within each of the soil zones (Agroecological Resource Areas) Results are presented as maps showing the distribution of av.ailable soil water at different levels of probabilities, for the most common AWC in each area. This information provides a fundamental understanding of the variability of available soil water across the different regions of the Canadian prairies, and the probability of a given level of growing season precipitation. This information reflects varying levels of production risk, and it can be used as an aid for decisions on farm practices, such as planting decisions summerfallowing or extended rotations. It can also be applied to policy and program decisions. . 3 METHODOLOGY Agroecological Resource Areas (ARAs) were employed to disaggregate the prairie region into relatively homogeneous biophysical land units. Plant available water was calculated for each ARA for the 1955-1985 period, and the results were subjected to a probability analysis. Further details on data, methods and assumptions follow. Agroecological Resource Areas (ARA) The agricultural portion of the prairies was divided into ARAs to provide a natural, soil landscape based framework for regional agricultural land evaluation. The criteria used to distinguish each ARA were based on agro-climate, surface form, soil texture and soil development. Each ARA was considered to be generally similar in terms of agricultural potential, land use and management (Pettapiece, 1989; Eilers and Mills, 1990; G. Padbury, personal communication) The dominant Available Water-holding Capacity (AWC) of each ARA was obtained by manually overlaying the ARA map with the maps published by De Jong and Shields (1988) Because the latter maps . often did not distinguish between soils having 50 or 100 mm AWC, these two classes were grouped for mapping purposes and assigned an AWC of 100 mm. Consequently four AWC classes were recognized, namely 100, 150, 200 and 250 mm, representing respectively sand, loam, clay loam, and clay soils (Fig. 2). ARA's dominated by Solonetzic soils, organic soils and high water tables were excluded from further analysis because the concept of available water as used herein did not apply to these soils. The AWC map of the ARA's (Fig. 2) sometimes did not coincide exactly with the maps published by De Jong and Shields (1988) because of scale differences in the base documents. Consequently, small polygons delineated by De Jong and Shields did not show up as dominant areas within an ARA. Moreover the ARA boundaries in Alberta did not always coincide with the polygon boundaries on the AWC maps of De Jong and Shields. Whenever this occurred, the AWC estimate was based on the largest area of the ARA falling within a given AWC of the base map. Some problems in mapping were experienced along the provincial boundaries. A large proportion of the soils in north-eastern Alberta were mapped as clay loams, whereas those across the border in Saskatchewan were mapped as loams. Similar, but less extensive problems occurred along the Saskatchewan/Manitoba border, but these were predominantly problems of local bias. No attempts were made to correct for boundary inconsistencies. The impact of this can be observed on some of the final maps, particularly if values in immediately adjacent areas on both sides of a border fell very near to the limits between two classes. . Climate Data Daily weather data from 1955 to 1985, including maximum and minimum air temperature, precipitation and potential evapotranspiration were derived for each ARA using the Thiessen polygon weighting technique (Williams and Hayhoe, 1982) The . technique was applied to a network of 165 climate stations for the 1955-65 period and 175 stations for the 1966-85 period. The full period was broken into these two segments in order to make maximum use of all available climate data. The weighting coefficients were checked and, when necessary, adjustments based on non- representative station elevations were made by local experts in each province. Modelling The modelling methodology (De Jong and Bootsma, 1988) used the Versatile Soil Moisture Budget (VSMB) (Baier et al., 1979) to estimate the components of the soil water balance for spring wheat. These procedures are based on the premise that water available for plant growth is gained by precipitation, but lost by evapotranspiration, runoff and deep drainage. The net loss or gain each day is added to the water already in the rooting zone of the soil. Water is withdrawn simultaneously, but at different rates, from different depths in the soil profile, depending on the rate of potential evapotranspiration, the stage of crop development, the water release characteristic of the soil and the available water content. The soil was subdivided into six standard zones, each having an available water-holding capacity calculated as a percentage of the total AWC (Baier et al., 1979). Water release characteristics were the same as those used by Sly (1982) Knowledge of seeding dates is essential for an accurate representation of crop growth and development. Observed data on "date when seeding is general" for spring wheat were obtained from Statistics Canada, Agriculture Division, Ottawa for each Crop Reporting District (CRD) for 1955-1985. The seeding date was then estimated for each ARA, by manually overlaying the ARA map with CRD- and land use maps (A. Mack, personal communication) and determining weighting factors based on the approximate fraction of cultivated land in each CRD located within an ARA. The rate of water uptake by the roots was simulated by crop coefficients which change as the crop goes through five phenological stages: seeding, emergence, jointing, heading, soft dough and maturity (or harvest) The duration of each growth stage . was defined by a biometeorological time scale model (Robertson, 1968) which required air temperature and photoperiod data. The , latter were calculated at the ARA's centroid latitude, using astronomical equations (Robertson and Russelo, 1968). No direct consideration was given to the effect of soil water conditions on phenological development.