Introduction

Predicting forage production for the upcoming growing season is an important aspect of rangeland and pasture management, specifically for proactive and adaptive grazing management decisions (Derner and Augustine 2016; Kelley et al. 2016). Climatic variability in intrinsically high interannual and intra-annual variable environments with substantial spatial variability in precipitation (Augustine 2010) magnifies rancher vulnerability to matching animal demand to forage availability (Eakin and Conley 2002; Scasta et al. 2015) and impacts economic returns (Bastain et al. 2012; Hamilton et al. 2016; Irisarri et al. 2019). Drought occurs about 30% of the time in the northern Great Plains (Smart et al. 2005). During drought, livestock producers are concerned with 1) having adequate forage for grazing livestock and hay harvest, 2) stockpiling forage for late fall/winter grazing, and 3) the persistence of current drought effects into the next growing season (Kachergis et al. 2014). These concerns emphasize the complexity of human dimensions of drought decision making (Wilmer and Fernández-Giménez 2015). This paper is a synthesis by rangeland management extension specialists and researchers attending a drought management symposium at the 72nd annual meeting of the Society for Range Management. We collected published, long-term forage production data sets from Alberta, Kansas, Montana, Nebraska, North Dakota, South Dakota, and Wyoming (Fig. 1 and Table 1) to demonstrate a broad, regional understanding of spring precipitation relationships to annual forage production, accumulated monthly forage growth (growth curves), and the significance and probability of relying on summer rainfall to compensate for spring drought in the central and northern Great Plains.

Precipitation Drives Forage Production

Precipitation is the primary driver of plant growth and soil moisture (Engda et al. 2016) and is typically the first limiting factor for forage growth in most grassland ecosystems of North America (Lauenroth and Sala 1992; Khumalo and Holechek 2005), including the Great Plains (Mowll et al. 2015). Additional factors, such as plant community (Smart et al. 2007; Derner et al. 2008), prior grazing history (stocking rate) (Launchbaugh 1967; Derner et al. 2007; Irisarri et al. 2016), time of grazing (Reece et al. 1996; Stephenson et al. 2015), topoedaphic position (Bork et al. 2001; Nippert et al. 2011; Stephenson et al. 2019), prior year precipitation (Petrie et al. 2018), and sea surface temperature anomalies (Chen et al. 2017) all have effects on forage production. For managers to make adaptive grazing management decisions for the upcoming grazing season, they need forage production predictions now available such as Grass-Cast (Peck et al. 2019) and remote sensing applications to handle both spatial and temporal relationships (Hermance et al. 2015; Gaffney et al. 2018); this information will assist managers in decision making related to drought (Dunn et al. 2005).

Figure 1.

The enclosed circles represent the locations of the climate diagrams in Figures 2 and 3 (USDI-USGS 2019).

Substantial precipitation and temperature gradients exist in the Great Plains (Lauenroth et al. 1994). Precipitation increases more than threefold from west to east (Burke et al. 1991; Anadón et al. 2014), and temperature increases from the northwest to the southeast (Epstein et al. 1996). Forage production increases nearly fourfold west to east (Heisler-White et al. 2009) due to precipitation (Lane et al. 1998) rather than with the temperature gradient (Epstein et al. 1996; Mowll et al. 2015).

In contrast to the precipitation-driven, large-scale, west-to-east influence on forage production (see earlier), smaller-scale (ranch-to county-level) forage production patterns are driven by soil water holding capacity (Sala et al. 1988), which combines local precipitation and soil characteristics. Foundational studies (Albertson and Weaver 1944; Dahl 1963) established this relationship at the smaller scales across several sites in the central Great Plains.

Precipitation and Plant Community Relationships

Growing season temperature and precipitation in the shortgrass prairie ecoregion have considerable overlap (Fig. 2), so forage production is robustly predicted using May–June precipitation (see Table 1). Seasonal distributions of precipitation and temperature have greater overlap in the mixed-grass prairie (see Figs. 2 and 3), with forage production predicted well by April–June precipitation across multiple locations in the northern and central Great Plains (see Table 1). The greater abundance of warm-season grasses in the tallgrass prairie shifts the influential months of precipitation to May–July for accurately predicting forage production (see Table 1).

Forage production of seeded perennial grasses and native rangeland vegetation display similar responses to seasonal precipitation. Variation in April and May precipitation explained > 90% of the variation in forage production of several seeded switchgrass (Panicum virgatum [L.]) cultivars, a native warm-season grass, in central South Dakota (Lee and Boe 2005). In eastern Montana, forage production of Russian wildrye (Psathyrostachys juncea [Fisch.] Nevski]), a non-native cool-season grass, was correlated with May and June precipitation (White 1985). Similar results were observed for the cool-season, non-native crested wheatgrass (Agropyron cristatum [L.] Gaertn.) in Colorado (Currie and Peterson 1966).

Climate change projections to 2050 have indicated that climate may trend greater temperatures, greater winter precipitation, and increased variability in growing season precipitation (Polley et al., 2013). These changes in climate may influence species composition in the central and northern Great Plains with more favorable conditions (e.g., earlier spring warm-up and increased winter precipitation) for cool-season grass production (Derner et al. 2018). Additionally, greater variability in precipitation events (e.g., more large and fewer small events) may affect total production differently based on grassland differences in the Great Plains. For example, growing season precipitation apportioned into a few large precipitation event sizes compared to many small event sizes increases aboveground net primary productivity in shortgrass and mixed-grass prairie (Heisler-White et al., 2009), but precipitation event size has less of an influence on primary production in tallgrass prairie systems (Wilcox et al., 2015). As such, long-term plant production data sets are critical for the continued tracking, evaluation, and understanding of potential changes in plant biomass shifts and how that may affect forage production growth patterns into the future. We hypothesize that precipitation received earlier in the year (e.g., March) will influence forage growth more in the future than it does now if, as predicted by Derner et al. (2018), warmer temperatures arrive earlier in spring and forage plants break winter dormancy earlier in the calendar year.

Table 1

Precipitation months that are important predictors and correlation coefficients (r) of annual forage production from grassland ecosystems in the central and northern Great Plains.

Figure 2.

Climate diagrams using 30-yr average (1981–2010) county data for Colorado, Montana, and Wyoming (HPRCC 2019). Growth curves are from ecological site descriptions (ESD) of the most common rangeland ecological site at each location (USDA-NRCS 2019).

Figure 3.

Climate diagrams using 30-yr average (1981–2010) county data for Kansas, Nebraska, North Dakota, and South Dakota (HPRCC 2019). Growth curves are from ecological site descriptions of the most common rangeland ecological site at each location (USDA-NRCS 2019).

Decision-Making Conundrums With Drought: The Need For Early Detection

Drought occurs quite regularly in the Great Plains as evidenced by the long-term datasets used in this analysis (Table 2). According to climate change projections for the region, the occurrence of drought is expected to increase (USGCRP 2018). Some regions in the Great Plains experience drought more often than others (see Table 2) making it especially important for managers and resources specialists to understand their local climate. Matching animal demand to forage availability becomes increasingly difficult as drought conditions emanate (Andales et al. 2006; Derner and Augustine 2016). Uncertainty about the likelihood of receiving precipitation and associated forage availability can delay making or implementing drought-related decisions (Haigh et al. 2019). Ranchers with limited flexibility and adaptive capacity in their operations (Kachergis et al. 2014) or without written drought plans with clearly identified trigger dates and associated predetermined decisions (Reece 2012) often make reactive rather than proactive decisions that have negative economic consequences (Dunn et al. 2005). These include reduced animal performance, purchasing supplemental feed, moving animals to a different location, and liquidating livestock during a depressed market because of high livestock supplies (Holechek 1996; Scasta et al. 2015; Scasta et al. 2016). The impacts of these negative consequences for ranching operations increase with increasing drought severity (Haigh et al. 2019). Delays in drought-related destocking influence economic returns and negatively impact rangeland productivity, health, and diversity (Haigh et al. 2019). The cumulative financial losses and mental stress from decision-making conundrums with drought impact emotional and social dynamics of ranching families and communities (Wilmer and Fernández-Giménez 2016).

Table 2

Frequency of occurrence of receiving < 75% of the mean (i.e., drought yr), 75–125% of the mean (i.e., average yr), and > 125% of the mean (i.e., wet yr) April–June precipitation and the percent of the median for the first and fourth quartiles at several locations in the central and northern Great Plains (HPRCC 2019).

Differences in resources and opportunities among ranching operations (locally and regionally) add to the complexity of determining which management strategies are best suited for adaptation to drought challenges (Wilmer et al. 2016). Each ranch has its own unique environmental, financial, and social conditions; therefore, optimal responses to drought vary among ranch businesses. As a result, the conceptual method of systems thinking, or examining linkages and interactions between different components of a ranch operation, may help producers better visualize, discuss, and apply drought management strategies for the whole operation rather than just being focused on drought alone (Rhoades et al. 2014). For example, increasing the proportion of stockers on a ranch may provide management flexibility and other beneficial opportunities (Kachergis et al. 2014). Dry spring conditions may allow producers to capitalize on enhanced spring individual stocker performance due to higher forage protein content (Harmoney and Jaeger 2011; Harmoney and Jaeger 2013; Owensby and Auen 2018) yet retain the ability to sell animals early in a drought to reduce forage demand. This strategy can provide more forage to the remaining cow herd and reduce the need to sell breeding stock. However, raising yearlings may require different marketing and management skills, and ranch goals and objectives may need to be refocused to include a yearling enterprise, thus altering other aspects of the whole ranch system. In an eastern New Mexico scenario, flexible grazing management with an equal forage allocation between cow-calf and yearling enterprises provided optimal opportunities for producers in this region to manage low and high forage production during dry and wet years, respectively (Torell et al. 2010).

Regardless of which strategies may be appropriate to help producers plan and adapt to drought conditions, early detection of drought is one of the most important variables in the decision-making process. Early drought detection can assist producers in making timely management choices and avoiding reactionary or forced decisions later in the growing season when fewer, and often less desirable, options are available. Producers can make early detection decisions by considering monthly forage growth expectations, long-term climate information, and short-term weather predictions. While other factors (e.g., livestock and hay markets) introduce uncertainty into decision making as well, stronger, clearer guidelines on forage production thresholds may help lessen this one key source of uncertainty, leading to increased confidence and more timely decision making.

Forage Growth Expectations

On the basis of the predictive relationships of spring precipitation and annual forage production (see Table 1), we used the growth curves of cool-season grasses and warm-season grasses grown in Hettinger, North Dakota and Ft. Pierre, South Dakota, respectively, as an example to understand the monthly forage growth expectation for a given year (Sedivec et al. 2009; Sedivec et al. 2010). Cool-season grasses, such as green needlegrass (Nassella viridula [Trin.] Barkworth), bluebunch wheatgrass (Pseudoroegneria spicata [Pursh] Á. Löve), and western wheatgrass (Pascopyrum smithii [Rydb.] Barkworth & D. R. Dewey), had a majority of their growth in June (48%) (Fig. 4a). Warm-season grasses, such as little bluestem (Schizachyrium scoparium [Michx.] Nash), blue grama (Bouteloua gracilis [Kunth] Lag. ex Griffiths), and sideoats grama (Bouteloua curtipendula [Michx.] Torr.), had a majority of their growth in July (36%) (see Fig. 4b). The Ecological Site Description (ESD) growth curves shown in Figures 2 and 3 were obtained from the US Department of Agriculture–Natural Resources Conservation Services (USDA-NRCS) (2019) ESD online database for 14 locations throughout the central and northern Great Plains based on the most common ecological site at each location. Growth curves are derived from a combination of field data and expert observation. The peak forage production month for a vast majority of the locations is June (see Figs. 2 and 3). Three exceptions to this were Havre, Montana; Miles City, Montana; and Broken Bow, Nebraska. The two locations in Montana had peak forage production in May, but June was important as well. The peak occurred in July for Broken Bow, Nebraska because the central mixed-grass prairie is dominated by warm-season grasses and has a longer growing season than northern mixed-grass prairie (see Fig. 3). The two locations in MT have a shorter growing season and a lower precipitation to temperature ratio in July and August, and they are dominated by cool-season grasses (see Fig. 2). In Montana, 75% of production occurs by early June and 90% by 1 July (see Table 2; Kruse et al. 2007; Vermeire et al. 2009). Adding water at a rate of 275% of the long-term median during July and August increased production following severe spring drought by < 2% of the median (Heitschmidt and Vermeire 2006). These data clearly indicate the importance of April–May precipitation and the futility of hoping for significant amounts of additional forage production in Montana after 1 July. In more central and eastern locations within the Great Plains, less of the annual production occurs by 1 July (Table 3), likely due to a longer growing season and a greater percentage composition of warm-season species (Stephenson et al. 2019).

Figure 4.

Growth curves for a, cool-season grasses and b, warm-season grasses grown at Hettinger, North Dakota and Ft. Pierre, South Dakota. In addition, the growth curve for a cool season–dominated ecological site description (ESD) and a warm season–dominated ESD were included (adapted from Sedivec et al. 2009; Sedivec et al. 2010; USDA-NRCS 2019).

We use contingent probabilities, or the likelihood of an outcome given the current conditions, to illustrate interpretations of the forage growth expectations for Miles City, Montana (Table 4). Here, we assume that 13 mm was received in April 2019; thus, 147 mm must be received in May + June to achieve the median 160 mm of total April + May + June precipitation. Over the past 30 yr, 147 mm or more precipitation in May + June was achieved 12 times (including a 146 mm amount in 1993). Therefore, a contingent probability of 40% chance (or 12/30) exists for median spring precipitation to total 160 mm if only 13 mm is received in April. Probabilities of receiving late spring or early summer precipitation to overcome spring deficits are greater in eastern portions of the Great Plains compared with western locations. Low probabilities of receiving the needed precipitation and the potential growth remaining after a dry spring period greatly increase the risk that average forage production will not be achieved after 1 July.

Table 3

Percent annual growth by 1 July and date of 90% annual growth for locations in the Great Plains. Data for growth curves were obtained from Ecological Site Descriptions (USDA-NRCS 2019).

Another way to analyze the probability of receiving late spring or summer precipitation to compensate for spring precipitation deficits is to use the Standardized Precipitation Index Explorer Tool ( https://uaclimateextension.shinyapps.io/SPItool/). The Standardized Precipitation Index (SPI) divides the difference between actual and median precipitation by the standard deviation of that distribution (McClaran and Wei 2014). SPI values of < –1 and > 1 each have a probability of occurring at ∼16%, and SPI values –1 to 0 and 0 to 1 each have probabilities of ∼34%. At Miles City, Montana (Table 5), there was never a case of a very wet July (> 1 SPI) after a very dry (< –1 SPI) April–June between 1980 and 2010. A wet (0 to 1 SPI) July occurred 60% of the time following a very dry (< minus 1 SPI) April-June. More important, however, is whether the 31 to 68 mm in a “wet” July is enough to pull out of a deficit of < 103 mm between April and June. The answer, of course, is “not likely” because artificial watering experiments showed little increase in forage production (Heitschmidt and Vermeire 2006).

Table 4

30-yr precipitation (mm) records for Miles City, Montana (HPRCC 2019). Cells highlighted are months and years where above-normal precipitation is needed to make up for spring moisture deficits outlined in the case scenario discussion.

Implications

Long-term studies of forage production have been used to develop predictive models using monthly precipitation. Our synthesis of the literature shows precipitation in April, May, and June (or some combination of these months) accurately predicted annual forage production in the central and northern Great Plains. This brief period at the beginning of the growing season represents an opportunity for ranchers and others to assess forage production and to make better informed drought management decisions ahead of market price signals. Forage growth curves for most of the central and northern Great Plains peak in June with a few locations peaking in July. This means that most (60–90%) of the annual forage production occurs by 1 July for cool season–dominated sites and 90% occurs by 1 September in warm season–dominated sites. The probability of receiving precipitation to make up for spring deficits decreases as the summer progresses. Ranchers and others can use short-term precipitation and forage prediction tools in the spring to evaluate the potential of developing drought conditions. By 1 July, drought coping strategies should be enacted or at least planned. While many ranchers wait until later in the summer or fall to respond to drought, drought response strategies should be enacted earlier. Waiting for July and August precipitation to make up for spring precipitation deficits to produce more forage is ill-advised because the probabilities are very low (usually < 20%) and the peak growth period has already occurred for most locations in the central and northern Great Plains. We suggest ranchers and others assess weekly precipitation, beginning in April each year, and make drought management preparations no later than late June or early July in the central and northern Great Plains.

Table 5

Standardized Precipitation Index (SPI) Explorer Tool ( https://uaclimateextension.shinyapps.io/SPItool/) showing the probabilities for transitioning from Period 1 (April–June) and Period 2 (July) for four categories (very dry, dry, wet, and very wet) for Miles City, Montana using the 30-yr period (1980–2010).

Figure 5.

Grass-Cast map produced on 29 April 2019 ( http://grasscast.agsci.colostate.edu/).