Actual and standard crop coefficients for semi-natural and planted grasslands and grasses: a review aimed at supporting water management to improve production and ecosystem services

Abstract

Natural and planted grasslands play a very important role in agriculture as source of various ecosystem services, including carbon sequestration and biodiversity, and are responsible for a large fraction of agricultural water use in rainfed and irrigated fields. It is, therefore, relevant to precisely know their water use and vegetation requirements with consideration of relevant climate, from extremely cold, dry, with long winter seasons, to tropical humid and hot climates, thus with a large variability of vegetation. Semi-natural grasslands are basically used for grazing and mainly refer to highland pastures and meadows, steppes, savannas, pampas, and mixed forest systems. The FAO method to compute crop (vegetation) evapotranspiration (ET_c) through the product of a crop coefficient (K_c) by the reference evapotranspiration (ET_o) is adopted. The selected papers were those where actual ET_c (ET_{c act}) was derived from field observations and ET_o was computed with the FAO56 definition, or with another method that could be referred to the former. Field derived ET_{c act} methods included soil water balance, Bowen ratio and eddy covariance measurements, as well as remote sensing vegetation indices or surface energy balance models, thus reviewed K_{c act} (ET_{c act}/ET_o) values were obtained from field data. These K_{c act} refer to initial, mid-season and end season (K_{c act ini}, K_{c act mid}, K_{c act end}) when reported values were daily or monthly; otherwise, only average values (K_{c act avg}) were collected. For cases relative to cold or freezing winters, data refer to the warm season only. For grasses cut for hay, K_{c act ini}, K_{c act mid}, and K_{c act end} refer to a cut cycle. K_{c act} values rarely exceeded 1.25, thus indicating that field measurements reported did respect the available energy for evaporation. Overall, K_{c act mid} for semi-natural grasslands in cold climates were lower than those in hot climates except when available water was high, with K_{c act mid} for meadows and mountain pastures generally high. Steppes have K_{c act mid} values lower than savannas. Grasses commonly planted for hay and for landscape generally showed high K_{c act mid} values, while a larger variability was observed with grasses for grazing. The collected K_{c act} values were used to define standard K_c values for all grassland and grasses. Nevertheless, the tabulated K_{c act} are indicative values of K_c to be used for actual water management purposes and/or irrigation scheduling of planted grasslands. It is expected that a better knowledge of the standard and/or indicative K_c values for a wide variety of grasslands and grasses will support better management aimed to improve grass productivity and ecosystem services, including biodiversity and carbon sequestration.

Introduction

Grassland is a main biome in Earth, occurring in every country or region, and having a great variability in relation to climate, landforms and elevation, environmental conditions, use and management. It includes rangelands, shrublands, pastureland, and cropland sown with pasture. Regional descriptions of grasslands of various types are provided in the FAO book edited by Suttie Reynolds Batello (2005). A recent classification and world mapping of grassland types, also referring to their biodiversity, has been provided by Dixon et al. (2014). The analysis by Seo (2021) reports many famed grasslands including the Pampas, the Llanos, the Prairie, the Steppe, the Savannas, and the Rangelands. The importance of grasslands is well recognized.

Grasslands account for 26% of the world’s global ice-free land area (Lorenz and Lal 2018), corresponding to the second greatest land use in the Earth (Lü et al. 2022), while forest land accounts for about 30%, and cultivated land for 12%. However, other authors considered that “grassland is the largest terrestrial biome on Earth” (Hobohm et al. 2021), accounting for up to 40% of the terrestrial area (Petermann and Buzhdygan 2021; Seo 2021). There are different concepts of grassland, e.g., with Dixon et al. (2014) defining grassland as a non-wetland type with at least 10% vegetation cover, dominated or co-dominated by graminoid and forb growth forms, and where the trees form a single-layer canopy with either less than 10% cover and 5 m height (temperate) or less than 40% cover and 8 m height (tropical). However, these limitations let out several semi-natural grasslands such as those of Mediterranean regions. Hobohm et al. (2021) state that grasslands support the livelihoods of 1 billion people with pastoralism (rising of livestock) with 20 million km² of grassland used for livestock feed production and for a variety of ecosystem services. Lü et al. (2022) pointed out the case of China, which has the second largest grassland area in the world, representing about 41% of China’s national territorial area. Mongolia has a smaller grassland area, but it corresponds to 83% of its land area (Angerer et al. 2008). The overall importance of grassland is therefore evidenced, which justifies the attention of many researchers to improve their use for livestock feed production and for ecosystem services.

The research studies relative to evapotranspiration (ET) cover many planted and semi-natural grasslands, as well as grasses, graminoids, and legumes used for planted grasslands and lawns. ET studies very often aim at assessing climate change impacts and future coping measures, or are relative to hydrologic and water resources assessments, particularly when referring to semi-natural grasslands. Studies focusing on management mostly refer to planted permanent grasslands for hay and grazing, namely when irrigated, and include water–grass–yield relationships. Research on grassland management focusing on ecosystem services is rare but many studies intend to recognize specific ecosystem services. It is well known that grasslands play an important role in ecological environment protection and animal husbandry development (Lü et al. 2022), but ecosystem services commonly identified often have relations with water despite these aspects are evaluated with a low value degree (Kang et al. 2020; Liu et al. 2022a). High valued ecosystem services refer to biodiversity, carbon sequestration, soil erosion control, and soil fertility enhancement, mainly when grass legumes are used. Runoff is retarded by the vegetation, thus favoring soil infiltration, which is also larger because grass cover impedes crust formation, therefore increasing water storage in the soil and improving water availability. Other commonly reported services include purifying chemical fertilizers and pesticides, and regulating groundwater, mainly in lowlands, while contributing to climate regulation, and extremes mitigation are also mentioned. Recreation, snow sports, and landscape aesthetics are extremely important ecosystems services in mountain areas of Europe and northern America, contrarily to other regions. Depending on the type and management of the grasslands, biodiversity and carbon fixation are quite relevant services, often object of research, namely in relation with the degradation of grasslands (Lal 2018; Hobohm et al. 2021).

According to Lal (2018), anthropogenic activities have affected about 40% of earth’s surface, and almost 92% of the natural grasslands and ecosystems, which have been converted to human use as grazing and croplands. Bonanomi et al. (2019) added that protecting forests at the expense of semi-natural grasslands can lead to the open-habitat loss of the Brazilian Cerrado biome. Hobohm et al. (2021) referred, as causes for degradation of grasslands, the expansion of urban areas, tree plantations, use of mineral fertilizers and pesticides, suppression of natural fires, over- and under-grazing, and intensification of use. Reforms of agri-environmental policies have aimed at incorporating environmental objectives into agriculture, such as biodiversity and carbon (C) sequestration in grasslands. However, many threats remain, “in both the now-fragmented areas of agriculturally improved productive lowlands” and in the marginal areas of Europe, where traditional systems are disappearing, and lands are abandoned (Hopkins and Holz 2006). Climate change, world population growth, and uncertainties over energy and water call for more focused research.

Land use conversion has depleted the terrestrial ecosystem C stock with major loss of the vegetation and soil C stock (Lal 2018; Lorenz and Lal 2018). Conversion to a restorative land use and adoption of good management practices may create a positive soil/ecosystem C budget that can lead to improved C sequestration rates in pastures, permanent crops, and lawns, and resulting from the restoration of soils prone to water erosion, also operated with grasslands. The adoption of best management practices—continuous ground cover, complex rotations, integrated nutrient management and no soil disturbance—can protect the soil organic carbon (SOC) stock and enhance ecosystem services (Lal 2018). Bai and Cotrufo (2022) reported that grasslands store near one-third of the global terrestrial C stocks, can act as an important soil carbon sink, with plant diversity increasing SOC storage.

Improved grazing management and biodiversity can provide C gains in global grasslands. Zhao et al. (2017) indicated that temperature, grazing intensity, and water availability are the major factors influencing SOC in grasslands of Inner Mongolia, China, while temperature and soil pH are more influencing in Mongolia, where grassland C sequestration is higher. Soussana et al. (2010) stated that soil carbon sequestration is the mechanism responsible for most of the greenhouse gas (GHG) mitigation potential in the agriculture sector and that grassland C sequestration has a strong potential to contribute for mitigating the GHG balance of ruminant production systems. However, CH₄ and N₂O emissions from livestock sector needs to be reduced and current SOC stocks preserved. More recently, Viglizzo et al. (2019) stated that grasslands sequester more carbon than forests because they are less sensitive to water stress and wildfires. This resilience of grasslands helps to preserve sequestered terrestrial C and prevent it from returning the atmosphere. The UN Decade on Ecosystem Restoration does not encourage afforestation of remaining semi-natural grassland and savannah ecosystems (Dudley et al. 2020) but proposes adopting a set of properly planned ecological, cultural, and social approaches for successful grassland and savannah restoration.

Enhancing biodiversity implies a good identification of grassland specialist species and of causes for favoring alien species richness. Noda et al. (2022) report that mowing is effective for the conservation of grassland specialists’ diversity, but it is required to pay attention to the invasion of alien species from adjacent areas. Biodiversity in rangelands is decreasing due to the intensification of their use for production (Alkemade et al. 2013). Extensively managed grasslands are recognized globally for their high biodiversity and their social and cultural values (Bengtsson et al. 2019). These authors propose that “ecosystem service and food security research and policy should give higher priority to understand how grasslands can be managed for fodder and meat production alongside other ecosystem service”. Texeira et al. (2015) reported that C gains are a key aspect of ecosystem functioning. In the Pampa biome, more than 80% of the species recorded by 1930 are still present, but the number of exotics has seven-fold increased (Burkart et al. 2011). In that case, the water availability was the main driving factor of floristic heterogeneity.

The brief review above definitely shows the importance of the grassland ecosystems at the world scale, as well as the importance of management for grassland to achieve improved production and ecosystem services, particularly C sequestration and biodiversity, and to mitigate and adapt to climate change. Grasslands management require knowledge of evapotranspiration (ET) as a main component of the water balance and as the driving force of plants transpiration and growth. Thus, considering the good number of published ET studies, it has been possible to perform a review aimed at extending the tabulated values of FAO56 (Allen et al. 1998), hence focusing on various types of grasslands, semi-natural ecosystems, and grasses. The review focused on the crop coefficient as defined by the FAO56 method (Allen et al. 1998), where vegetation (crop) ET (ET_c) is computed as ET_c = K_c ET_o, product of the reference ET (ET_o), also known as potential ET (PET, assumed equal to ET_o), by the specific (vegetation) crop coefficient (K_c). Thus, the articles published after 1998 were targeted. The FAO56 method (Allen et al. 1998) was the most often used in the papers reviewed, is the most common and easy method used for the generality of agricultural crops in the field practice, and their K_c are tabulated in FAO56. For these reasons, and as an opportunity to update and expand the Tables in FAO56, the FAO method was selected for the current review.

Nevertheless, other approaches to compute ET_c were adopted in the reviewed studies, which also computed ET_o and actual (ET_{c act}), thus allowing to obtain K_{c act} = ET_{c act}/ET_o. Therefore, the current paper shows the tabulated values of K_{c act} for irrigated and non-irrigated grasslands, meadows, and pastures, for semi-natural vegetation consisting of steppes, savannas and other ecosystems, and tabulated K_{c act} values for grasses, with distinction of their use for animal production or for landscape, presented in Sections “Seminatural and planted grasslands”, “Semi-natural grassland ecosystems” and “Grasses for hay, grazing and landscape”, respectively. The analysis of the tabulated K_{c act} allowed to derive standard transferable K_c values for the considered grassland ecosystems and grasses, which are presented in Section “Standard crop coefficients”. Therefore, the current review consists of a full update and extension of standard K_c values proposed in the FAO56 guidelines for computing crop evapotranspiration aimed at supporting improved field and water management of grasslands, and more accurate water balances and hydrologic studies, thus easing the consideration of water balances in studies relative to ecosystem services where water plays a role.

Materials and methods

The review aimed at collecting the available K_{c act} for grasslands and was performed through the widest possible search focused on papers reporting on actual K_c obtained from field measurements of grasses and grasslands actual evapotranspiration (ET_{c act}).

The search was performed in Science Direct and through the on-line pages of various journals, as well as using the bibliography lists of selected articles. Several languages were considered: English, Spanish, Portuguese, French, Italian, and German. In addition to the keywords evapotranspiration and crop coefficients, numerous other keywords were used including grass, semi-natural grasslands, planted grasslands, pastures, meadows, rangelands, steppe, savanna, prairie, shrubland, pampas, chaparral, and páramos. Only full articles were reviewed.

The criteria used for the selection of the papers from where single and basal K_{c act} values were collected consist of the following:

• (1) The papers should be of good/acceptable quality, without preference of journals where published.

• (2) The K_{c act} values should have been derived from adequate field research, exceptionally from solid review papers.

• (3) The field methods should be well described and readable by any interested reader, and should refer to consistent methodologies that provide for computing the ET_{c act}, including when less common empirical field methods were used.

• (4) The grass reference ET_o should be computed with full daily data sets using the FAO Penman Monteith equation (FAO-PM-ET_o). When a different equation was used, including when data sets with missing variables were available, either the ratio of the equation used to the FAO-PM-ET_o was commonly known, or information was available from the authors; a conversion factor of 1.15 was used when the ASCE-PM ET_r equation for alfalfa (Allen et al. 2006) or the Penman equation (Doorenbos and Pruitt 1977) were adopted.

• (5) Under the conditions referred before, the K_{c act} values were provided by the authors in Tables, graphics or in the text; for a few cases, when only ET_{c act} and ET_o were provided, K_{c act} (average) values were computed. Otherwise, data could not be considered.

• (6) Another important aspect was relative to the description of the studied grassland; when information was too brief the paper was excluded; nevertheless, depending upon the rarity of the crop’s information, data from papers where that information was less good, namely on botanical data, were used.

• (7) In addition, the field and computation methods should be sufficiently descriptive, in line with the recommendations by Allen et al. (2011), to understand if the reported methods provided for reliable ET_{c act} data. Otherwise, the study was not considered.

Among research studies on grasslands ET, many did not use the FAO56 method but, as for the generality of other methods, required specific field procedures. Field research methods included: the soil–water balance (SWB) based on observations of the soil water content (SWC) using soil sampling and various types of sensors; the field or catchment hydrologic water balance (HWB); the Bowen ratio energy balance (BREB); the eddy covariance system (EC); weighing and drainage lysimeters (WL and DL); mini or micro lysimeters (ML) to assess soil evaporation; and diverse but consistent empirical methods such as testing different K_c values against observed yields. Most field methods are described and analyzed for accuracy by Allen et al. (2011).

The methods used to compute and assess ET_{c act}, in addition to the FAO56 method (Allen et al. 1998), included the Penman method (Doorenbos and Pruitt 1977), the Penman-Monteith combination equation (Montheith 1965; PM), the Priestley-Taylor equation (Priestley and Taylor 1972; PT), and the double source method of Shuttleworth and Wallace (Shuttleworth and Wallace 1985; SW). The PM, the PT and the SW equations require specific field methods different from the SWB. Several studies were performed with support of properly calibrated models. The most used software models comprise SIMDualKc (Rosa et al. 2012; Pereira et al. 2020) and HYDRUS (Šimůnek et al. 2016). In addition, remote sensing (RS) was largely used mainly in the last decade. Both surface energy balance models such as METRIC, SEBAL, and SEBS (Allen et al. 2007), and RS vegetation indices such as NDVI (Glenn et al. 2011; Pôças et al. 2020), were adopted. A list of symbols, acronyms and abbreviations is included in Appendix B.

The tables for grasslands K_{c act} were divided into three main groups with each one divided again according to the types of grasslands, and where the reviewed papers are “grouped” following the ecosystems type and/or farm use:

• 1. Semi-natural and planted grasslands and meadows, divided into semi-natural, non-irrigated and irrigated,

• 2. Semi-natural grassland ecosystems, comprising savannas, steppes, and other semi-natural ecosystems such as mixed forests and shrublands, and

• 3. Grasses for hay, for grazing and for landscape uses.

The tables provide information on the location of the studied sites, the authors, and the main grasses and the actual crop coefficients, and the conditions corresponding to the determination of the presented actual K_c values for the initial, mid-season and end season (K_{c ini}, K_{c mid} and K_c end respectively, Fig. 1), following the FAO56 definitions (Allen et al. 1998). The information provided aims at easing the transferability of K_c values, including when users consult the original papers.

Fig. 1

Typical K_c/K_cb curves for a grassland used for grazing or seed (FAO56, Allen et al. 1998)

Information on climate, the methods used for determining ET_o and ET_{c act}, the management adopted, the growing season and water supply are given in Appendix A, complementing data in Tables 1, 2, and 3.

Table 1 Field derived actual crop coefficients (K_{c act}) for semi-natural high mountain grasslands and meadows for grazing and hay, with identification of the study sites, article reference, and main grasses

Table 2 Field derived actual crop coefficients (K_{c act} and K_{cb act}) for non-irrigated grasslands and meadows in highland plateau and prairies

Table 3 Field derived actual single and basal crop coefficients (K_{c act} and K_{cb act}) for irrigated grasslands, meadows, and pastures

The selected papers designate grasses, shrubs, and trees with the scientific or the common names. In this review paper, the scientific names are used. To ease recognizing the plants, a Table in Appendix C lists the scientific names used and the corresponding common name when known.

The tabulated actual K_c values are not adjusted to climate (Allen et al. 1998) because grasslands very often have a small height, thus small variation of K_c with wind speed and relative humidity. In addition, users generally know well that the transferability of K_c information is related to the type of grassland and the environmental conditions and, therefore, they are able to well transfer the tabulated values adopting small empirical corrections to K_c, of 5–10%, which values may slightly increase K_c in dry and windy conditions, and lower K_c in humid and calm environments as recommended in FAO56 (Allen et al. 1998).

The dual K_c approach proposed in FAO56 (Fig. 1a) was used by a few authors who reported basal crop coefficients (K_cb) representing the transpiration (T_c) component of ET, i.e., K_cb = T_c/ET_o (Allen et al. 1998; Pereira et al. 2020). The actual K_cb values are tabulated together with the single K_{c act} but using a bold and italic format of the characters. Since grasslands typically have a large density of plants well shading the ground, it results in a small soil evaporation (Paredes et al. 2018) and quite small differences to the single K_c, thus concluding that information on single averaged K_c is definitely sufficient for further assessment of water use and water balance.

Seminatural and planted grasslands

This section refers to actual K_c for the various types of grasslands that are used for animal feeding through grazing and hay, or to produce grass seed. The generic grasslands name is used. The first group of grasslands consist of the irrigated and non-irrigated grasslands, the latter being the most common ones, that include numerous semi-natural grasslands, which water supply is precipitation or, less often, high groundwater tables.

Semi-natural and planted non-irrigated grasslands

Table 1 shows K_c act for semi-natural grasslands, meadows, and pastures in high elevation sites with identification of the field study location, reference of selected articles, the floristic composition, and conditions relative to the crop when the K_{c act} were derived. Semi-natural grasslands occur in various cold and temperate climate ecosystems, as meadows in high elevation mountains of the Qilian Mountains and the Tibetan Plateau, the Andean páramos of the Equator, or the Alpine pasture of the Aosta Valley (Table 1).

The highland sites show large K_{c act} values, around 1.0, for the summer, unfreeze period, similar to the high elevation grasslands reported in Section “Semi-natural and planted non-irrigated grasslands”. Those high values mean that soil water was well available for satisfaction of the vegetation after the winter snow and ice melting. The period of grass growth has both high water and energy availability, however, with growth limited by the temperature. In high mountain and plateau locations, the winter is long and freezing, so reducing the mid-season and conditions for killing frost exist causing that the end-season may be anticipated. They are located far from farms and the rural population. The high elevation grasslands are generally semi-natural, whose reported grasses are rarely planted (not included in the Tables for grasses in Section “Grasses for hay, grazing and landscape”).

Table 2 reports K_{c act} for low mountain and lowland grasslands, where the crop season is longer, the winter is less cold and grasslands are located not far from farms, thus where human interventions have occurred, altering the flora by planting more productive grasses, and using fertilizers and pesticides, and adopting other management practices, like cutting for hay, not usual in semi-natural grasslands. The low mountain non-irrigated grasslands may be semi-natural, but those in low land are very often planted. The abandonment or mismanagement of the semi-natural grasslands also cause alterations due to the progressive invasion of shrubs and trees, which compete with grasses for water, nutrients, and energy.

The reported values (Table 1) show that K_{c act ini} range 0.25 to 0.65, corresponding to the regrowth of grasses in spring, that depends upon the soil water availability by then. They are much lower than the K_c act mid, which ranges from 0.85 to 1.20; there are, however, lower mid-season values when water is much insufficient. K_{c act end} values may be quite lower than K_{c mid} when solar radiation progressively decreases, then followed by the decrease and end of growth, particularly for higher latitudes, and cold time installs. End season values may be close to the mid-season when growth stops abruptly due to sudden changes of temperature and radiation in the Fall due to killing frost occurrence. This condition is not likely to occur in less elevation grasslands (Table 1).

The K_{c act mid} tend to be higher where water is available, either due to rainfall or, in altitude, due to snow and ice melt; lower values refer to the grasslands in plateau steppes (e.g., Xilin, Zhao et al. 2010), where water availability is scarce. For the cases where only an average K_{c act} value is reported, it may be observed that higher actual K_{c avg} correspond to conditions similar to high K_{c act mid}. Overall, small values are for grasslands in dry areas.

Selected main characteristics of the grasslands description, namely the crop season period, water supply (precipitation and/or groundwater (GW)) and the methods used for determination of ET_o and ET_{c act}, are provided in Appendix A, in Tables 12 and 13, in correspondence to the Tables 1 and 2. Analyzing the variability of K_{c act} values, it was found that those data are important to identify the type of grasslands and the quality of field and lab research work behind the reported K_{c act}, but do not help to explain their variability.

Irrigated grasslands

Irrigated grasslands are commonly located in lowland areas or in low slope fields, generally of low altitude, and are planted for grazing or for hay. The information on sites and K_c act and K_{cb act} values are reported in Table 3, while data further characterizing the irrigated grasslands are shown in Appendix (Table 14).

When grasslands are cropped for hay, their K_{c act} values refer to the observed cycles of grass cutting or to an average or representative cycle according to decision criteria of authors. Each cycle is described by the common FAO K_c curve (Fig. 1) comprising four crop coefficient stages—initial, development, mid-season, and late season stages—as presented in Fig. 2 for a case with four cycles. Thus, each cycle is characterized by a K_c ini, K_{c mid} and K_{c end}. The K_{c ini} corresponds to the K_c that follows the cut, while K_{c end} refers to the K_c when the cut is performed, the K_{c cut}. Grass height h and cover fraction f_c may use the same subscripts as for K_c.

Fig. 2

Typical K_c/K_cb curves for a grassland cropped for hay with multiple cutting cycles (FAO56, Allen et al. 1998)

In rotary grazing, similarly, there are various cycles comprising a period when the livestock is grazing followed by a period of grass development until animals start grazing again. Only two K_c are necessary for fully describing these cycles, the K_{c high} when the animals enter in the grass field, and K_{c low} when they end grazing. For turf grass in any landscape grass is mowed to the height h_low when it attains the height h_max. Researchers, however, do not yet adopt a standardized nomenclature, which may result in confusing. This nomenclature is used in the Tables presented in this article.

Table 3 shows that, with a single exception, the reported K_{c act avg} and the K_{c act mid} for irrigated grasslands are close to 1.0 with values ranging between 0.80 and 1.20. These values are higher than for non-irrigated grasslands (Tables 1 and 2) because there is more water available due to irrigation and, likely, the grasses’ soils are often of better quality and management of nutrients are more careful. An exception refers to a deficit irrigation site in the arid Gareh Bigorn Plain, southern Iran (Pakparvar et al. 2014). The reported ground cover fraction is always high, i.e., f_c = 1.0, which indicates that grasslands were well managed. Most of the cases in Table 3 report average information; only one case refers 3 cycles cutting for hay.

Two cases report K_{cb act} values, one with K_{cb act mid} lower than K_{cb act ini} and K_{cb act end} because the site is dry and hot (Imperial Valley, California, Allen et al. 2005a), thus affecting plant growth, so with higher values when weather is mild. The other (in Victoria, Australia, Greenwood et al. 2009) reports experimental K_{cb act} results that correspond to good plant growth, thus to high K_{cb act}.

Semi-natural grassland ecosystems

Semi-natural savanna and steppe type grasslands

Both savanna and steppe designations, herein, do not refer to specific biomes but include a variety of other biome like cerrado and catinga in Brazil, dehesa or montado in the Iberian Peninsula, or to grasslands in open forests. Naturally, the grasslands included under those designations vary much with climate and, regionally, with the dominant species and environment, particularly with soils. Various savanna-type semi-natural grasslands are reported in Table 4. Despite these grasslands are used for grazing after long time, the grasses and shrubs are different from those in the planted grasslands and of the domesticated grasses reported in Section “Grasses for hay, grazing and landscape” hereafter. Because savanna grasslands are not irrigated, both K_{c act mid} and K_{c act avg} show a seasonal effect related with the precipitation regime, with higher K_{c act} in the rainy seasons, not when more solar energy is available, i.e., contrarily to reported grasslands in the preceding section, savannas are mostly water limited and less energy limited. The case of the oak savanna (montado) of Évora (Paço et al. 2009), showing very low K_{c act} due to severe drought, is a good example. A unique example of effects of savanna conservation is provided by Descheemaeker et al. (2009, 2011), where grazed and protected savanna show different K_{c act mid} values, higher in the latter case due to better growth of the vegetation.

Table 4 Field derived actual crop coefficients (K_{c act}) for semi-natural savanna-type grasslands

In case of steppe (Table 5), there are similar behaviors, as for the high plateau steppe in Inner Mongolia reported by Zhang et al. (2012), and for two Brazilian catinga studies (Teixeira 2010; Carvalho et al. 2018), all referring higher K_{c act} values when it rains and the soil water availability increases. There are various cases where K_{c act} for protected or well managed steppe grasslands is much higher than for commonly grazed steppes, e.g., Miao et al. (2009) and Lu et al. (2011) relative to the high plateau of Inner Mongolia. K_{c act} results for steppe, like for savanna, indicate that related grasslands plant development is mainly water limited and less energy limited. This fact is important for management and relative to ecosystem services; therefore, in agreement with K_c act results analyzed in the previous section, it allows to consider that water management of grasslands may have implication on various services, mainly biodiversity and carbon sequestration, since these services are better when plants grow favorably.

Table 5 Field derived actual crop coefficients (K_{c act}) for semi-natural steppe grasslands

Semi-natural grasslands in cold and temperate ecosystems

The pampa grasslands, at a low altitude, show K_{c act avg} near 0.85 without evident distinction between seasons, likely due to a more favorable precipitation regime. Semi-natural grasslands in low precipitation areas have a lower K_{c act mid} or K_{c act avg} than pampa sites and show the seasonal influence of the rainfall regime. It is important to note that main grasses in Table 6 are different from a site to another.

Table 6 Field derived actual crop coefficients (K_c) for semi-natural grasslands in temperate climates

Semi-natural grasslands in mixed forests and shrublands

Grazing is common in open mixed forests where grasses are often native if management did not favor the loss of semi-natural grass vegetation in favor of alien species. Contrarily, in planted forests, it is common that the native/semi-native understory vegetation has changed after introducing the new tree species. It is, therefore, likely that grasslands growing as understory of mixed forests area are considered semi-natural.

Table 7 shows various sites where this condition could be accepted but which research papers may have not provided related full information. Data in Table 7 show that both dry and humid climates, e.g., Roupsard et al. (2006) and Corbari et al. (2017), have K_{c act} varying seasonally in relation to water availability. In general, K_{c act} of mixed forests varies in a small range, 0.45 to 0.60. Shrublands show higher K_{c act} values than mixed forests, likely because shrub roots can explore the soil to a large depth and solar energy available to grass is less affected by shadow, so overall contributing to a higher actual K_c.

Table 7 Field derived actual crop coefficients (K_{c act}) for semi-natural grasslands in mixed forests and shrublands

Grasses for hay, grazing and landscape

This Section “Grasses for hay, grazing and landscape” refers to domesticated grasses used in agricultural planted grasslands and in landscape and sport fields, which are described in Tables 8, 9 and 10. It may be noted that these domesticated grasses were rarely reported among the main grasses of semi-natural grasslands, in previous Tables 1, 2, 3, 4, 5, 6 and 7.

Table 8 Field derived actual single and dual crop coefficients (K_{c act} and K_{cb act}) for grasses used for hay or grazing with multiple cut cycles

Table 9 Field derived actual single and dual crop coefficients (K_{c act} and K_{cb act}) for grasses cropped for grazing and seed production

Table 10 Field derived actual single crop coefficients (K_c) for landscape grasses

Grasses for hay are mainly legume-grasses that grow fast under favorable environmental conditions and that respond well to cuts and allow numerous cut cycles during a crop season as represented in Fig. 2. For most cases, tabulated actual K_{c ini}, K_{c mid} and K_{c end} values describe the cut cycles; otherwise, only K_{c act avg} was reported for one grass.

Alfalfa is the most common grass for hay and the most studied one, namely with four papers using the dual K_c approach (Table 8). Results are quite similar, with actual K_{cb ini}, K_{cb mid} and K_{cb end} of approximately 0.30, 1.15 and 1.10, respectively. The higher mid-season value, reported by Hunsaker et al. (2002), shows the effect of a dry, hot, and windy climate. The reported values, considering that alfalfa grass covers well the soil (f_c~1.0), result in K_{c act} quite close to K_{cb act}, thus, actual K_{c ini}, K_{c mid} and K_c end of 0.40, 1.20 and 1.15, respectively. These values are coherent when compared to the standard tabulated values in FAO56 (Allen et al. 1998) and consist of standard K_c.

Several grasses have K_{cb act} values similar to those of alfalfa (Table 8). However, most of them show K_c act values varying with the cuts due to seasonality effects, which relate with climate dryness or wetness and windy conditions, more important when the grass is high by the mid and end stages, as proposed in the FAO56 equation for correction with climate. This is typically the case for blue panic cropped at Jeddah, Saudi Arabia (Ismail and El-Nakhlawy 2018). To be also noted that end-season K_c act maybe larger or equal then K_{c act} at mid-season, despite it is commonly a little smaller for most cases. The reported value is only K_c avg in case of palisade grass (Antoniel et al. 2016), which corresponds to a less accurate field measurement but quite useful to indicate that this grass (Brachiaria brizantha) likely is a high-water demanding crop. However, information in this Table 8 is more useful when users compare various grasses.

Table 9 refers to the field derived actual single and dual crop coefficients for grasses cropped for grazing and seed production. Often, only one K_c/K_cb curve is required. However, a precise management requires that specific curves (Fig. 2) are used in rotary grazing, when h_low and h_cut are well defined, or when the time interval between cuts is defined with the cumulative growth degree days (CGDD). The latter is shown through an example with Bermuda grass cv “Tifton 85” experimentally cropped in southern Brazil (Paredes et al. 2018). This case also shows that for small grazing time intervals there is no need to adopt grazing cycles (Fig. 2), but this becomes of interest when such intervals between successive grazing events are large and differences between crop heights and K_c are larger, e.g., if CGDD=372 °C is adopted (Paredes et al. 2018).

It may be advisable to adopt different K_c values for groundwater-fed grass, where K_{c act} varies with the water table depth (WTD), e.g., the Timothy and Italian ryegrass cases referred by Mueller et al. (2005). Generally, K_{c act mid} varies in the range 0.80 to 1.00 but K_{c act ini} and K_{c act end} values have a larger range of variation, which is likely due to management and to climate, mainly relative humidity of the air and wind speed. Grasses cropped for seed have smaller K_{c act end} since they are harvested following senescence and maturation of the seeds.

The grasses used for landscape (Table 10) are those able to live healthy and fully covering the ground while being frequently or very frequently mowed to small (5–8 cm) or very small heights (< 1.5 cm) as used respectively for lawns and for golf courses. Generally, knowing a single K_{c avg} is enough for a good irrigation, commonly in the range 0.60–0.80 for lawns and larger in case of golf greens because the requirements of quality are much larger for the latter.

The grass actual K_c values summarized in Tables 8, 9 and 10 concern grass fields with large f_c (> 0.95), and they are appropriate for computing ET for in hydrologic and water resources studies.

Standard crop coefficients

From the analysis above and taking into consideration the tabulated information (Tables 1, 2, 3, 4, 5, 6 and 7) and the related papers, it is possible to propose a set of standard K_c values for the referred grasslands. Nevertheless, the previously tabulated actual K_c values may be used as indicative values for management or planning, e.g., for use to estimate ET in irrigation scheduling tools or models applied only to similar grasslands, i.e., not generally transferable. Differently, the standard K_c values, to be tabulated in FAO56 and shown in the Table 11, are transferable for a wider use relative to the corresponding types of grasslands, i.e., in irrigation scheduling tools and models and in hydrologic or water resources studies and models. Particular attention must be given to the climate, comparing conditions in the original location, summarily indicated in the Tables, and in the location where the transferred K_c is to be used. The defined standard K_c values for semi-natural and planted grasslands and for grasses for animal feeding and landscape uses are presented in Table 11.

Table 11 Standard K_c values for semi-natural and planted grasslands and for grasses for agricultural and landscape uses

The proper use of standard values of grasslands implies that user knows that tabulated values refer to non-stressed or mild-stressed vegetation. Tables 1, 2, 3, 4, 5, 6 and 7 show that low K_{c act} values occur often, particularly for semi-natural vegetation in dry climates, namely steppe and savanna ecosystems, where actual K_c may vary much. Thus, when wishing to transfer a K_c to a dry or a drought prone area it is advisable to pay attention to the tabulated actual K_c (Tables 1, 2, 3, 4, 5, 6 and 7). The same happens with the use of standard values of grasses (Tables 8, 9 and 10). Their tabulated values are generally non-stressed or mild stressed. It is our conviction that for both grasslands and grasses transferability is adequate if users analyze carefully all related Tables and, in addition to climate, also take into consideration the dominant species.

Conclusions

The current review presents to users a large information on crop coefficients for determining crop evapotranspiration and, thus, to support new approaches for management taking into consideration both production for animal feeding and ecosystem services. Moreover, the review has shown that a large fraction of the grasslands is semi-natural and, therefore, may help in fighting climate change if appropriately managed for conservation.

The first group of grasslands focused those that are being used for grazing or hay, planted or semi-natural, normally using mixed grasses. The majority are non-irrigated and include a good number of semi-natural mountain pastures and meadows. Their growth conditions are linked to water availability, thus showing a wide range of actual K_{c mid} and K_c avg values. Despite management is not referred to water but rarely, this group of papers (section “Seminatural and planted grasslands”) makes it somewhat evident that ecosystem services, such like biodiversity, C sequestration, and runoff and erosion control call for more importance to be given to water use in grasslands management.

The second group (Section “Semi-natural grassland ecosystems”) refers to grasslands in various typical biome, covering a wide range of environments and ecosystems, from hot and dry plains to freezing and humid mountainous areas. These types of grasses helped to identify the need for consideration of water in management of such varied types of semi-natural grasslands and to associate water and grazing management to avoid grassland deterioration and to provide for biodiversity and C sequestration.

A variety of grasses for most of environments and grassland uses are described in Section “Grasses for hay, grazing and landscape”. Since they are used as planted grasses, related information is important for new plantations, using both single and combined grasses. Moreover, that information is useful for irrigation management and scheduling applied to irrigated grasslands. Related applied research should be developed aiming at improved water productivity and water saving since such information is rare.

K_c values tabulated for that wide number of grasslands and grasses may be useful for feeding all kind of herbivorous, for landscape and for sport activities, always considering the need for saving water, i.e., to avoid excess water application and, on the contrary, to avoid detrimental water deficits that reduce both the productivity and the ecosystem services. It is opportune to refer the need for continuing research that may not only increase the transferable case studies data but also may support improving the summarized standard K_c values (Section “Standard crop coefficients”) for use in Hydrology and water resources. This review led to conclude that research on grass productivity should also consider issues for ecosystem services.

Research aimed at ecosystem services requires however a better consideration of the role that water plays to improve biodiversity, C sequestration, water infiltration, thus controlling runoff and erosion, improving water availability through storage in the soil and in groundwater, thus contributing to mitigate effects of climate extremes and climate change, particularly in case of semi-natural grasslands. More research is required along these lines as well as relative to policy making that could contribute to define related priorities and the protection of semi-natural grasslands, as well supporting the mitigation of impacts of global change.

Appendices

Appendix A

See Tables 12, 13 and 14.

Table 12 Characteristics of selected semi-natural high elevation grasslands.

Table 13 Characteristics of selected non-irrigated grasslands in low elevation mountains and lowlands.

Table 14 Field observed K_c and K_cb for irrigated grasslands, meadows, and pastures

Appendix B. List of symbols, abbreviations, and acronyms

ET_c:

Crop evapotranspiration under standard conditions [mm d^-1 or mm h^-1]

ET_{c act}:

Actual crop evapotranspiration, i.e., under non-standard conditions [mm d^-1 or mm h^-1]

ET_o:

(grass) reference crop evapotranspiration [mm d^-1 or mm h^-1]

ET_r:

Alfalfa reference crop evapotranspiration [mm d^-1 or mm h^-1]

f_c:

Fraction of soil surface covered by vegetation (as observed from overhead) [-]

h_cut:

Crop height before cutting [m]

h_low:

Crop height after mowing or cutting [m]

h_max:

Crop height before mowing or grazing [m]

K_c:

(standard) crop coefficient [-]

K_{c act}:

Actual crop coefficient (under non-standard conditions) [-]

K_{c avg}:

(standard) average crop coefficient [-]

K_{c ini}:

Crop coefficient during the initial growth stage [-]

K_{c mid}:

Crop coefficient during the mid-season growth stage [-]

K_{c end}:

Crop coefficient at end of the late season growth stage [-]

K_{c cut}:

Crop coefficient before cutting [-]

K_{c high}:

Crop coefficient prior to grazing starts [-]

K_{c low}:

Crop coefficient at the end of grazing [-]

K_cb:

Standard basal crop coefficient [-]

K_{cb act}:

Actual basal crop coefficient (under non-standard conditions and/or observed) [-]

K_{cb ini}:

Basal crop coefficient during the initial growth stage [-]

K_{cb mid}:

Basal crop coefficient during the mid-season growth stage [-]

K_{cb end}:

Basal crop coefficient at end of the late season growth stage [-]

K_s:

Water stress coefficient [-]

PET:

Potential evapotranspiration [mm d^-1 or mm h^-1]

T_c:

Crop transpiration [mm d^-1 or mm h^-1]

Abbreviations and acronyms

ASCE-PM-ET_r:

Alfalfa reference ET_r calculated using an extension of the FAO56 Penman-Monteith equation

Avg.:

Average

BREB:

Bowen ratio energy balance

Capacit.:

Capacitance sensors

CGDD:

Cumulative growing degree day [^oC]

DL:

Drainage lysimeters

EC:

Eddy covariance

ECV-SM:

European Space Agency and Climate Change Initiative merged soil moisture product

EVI:

Enhanced Vegetation Index

FAO:

Food and Agriculture Organization

FAO56:

Food and Agriculture Organization Irrigation and Drainage Paper 56 (1998)

FAO56-PM-ET_o:

Grass reference ET_o computed with the FAO56 standardized Penman-Monteith equation

FLUXNET:

Global network of micrometeorological flux measurement sites

GHG:

Greenhouse gas

Grav.:

Gravimetric method

GW:

Groundwater

GW Lys.:

Water table lysimeter

HWB:

Field or catchment hydrologic water balance

J&H:

Jensen and Haise equation

LAI:

leaf area index

Med:

Mediterranean

METRIC:

Energy Balance model for Mapping EvapoTranspiration with Internalized Calibration

ML:

Mini or micro lysimeters

MODIS:

Moderate Resolution Imaging Spectroradiometer

NDVI:

Normalized Difference Vegetation Index

PM-eq.:

Penman-Monteith combination equation

PT:

Priestley-Taylor equation

Reflect.:

Reflectometer

RS:

Remote sensing

SAFER:

Simple Algorithm for Evapotranspiration Retrieving

SAVI:

Soil adjusted vegetation Index

SEB:

Surface energy balance

SEBAL:

Surface Energy Balance Algorithm for Land model

SEBS:

Surface Energy Balance System model

SF:

Sap flow

SOC:

Soil organic carbon

Spg:

Spring

Spr:

Sprinkler

SR:

Surface renewal

Sum.:

Summer

SW:

Double source method of Shuttleworth and Wallace

SWB:

Soil water balance

SWC:

Soil water content

Tens.:

Tensiometers

Trime:

Trime-EZ soil moisture sensors

UN:

United Nations

VI:

Vegetation index

Win:

Winter

WL:

Weighing lysimeter

Appendix C. Scientific and common names of the plants mentioned in the previous Tables

Scientific name	Common name	Scientific name	Common name
Acacia spp.	Wattle, mimosa, thorntee	Festuca rubra	Creeping red fescue
Acacia etbaica	Clownhair wattle	Festuca spp.	Fescue grass
Acacia senegal	Gum Acacia, Gum Arabic Tree, or Gum Senegal Tree	Foeniculum vulgare	Common fennel
Acacia victoria	Gundabluie, or bardi bush	Geoffroea spp.	Chanar, Chilean Palo Verde
Achnatherum sibiricum	Siberian Needlegrass	Geranium spp.	Cranesbills
Aegilops crassa	Persian goatgrass	Haloxylon ammodendron	Saxaul
Aristida affinis = A. purpurascens	Arrowfeather threeawn	Helianthemum lippii	Raqrouq
Aristida laevis	Aristida grass	Holcus lanatus	Yorkshire fog, fog grass
Agrostis spp.	Bentgrass	Iriantus angustifolium
Agrostis stolonifera	Bentgrass, creeping bent	Kobresia sp	Perennial sedge.
Alisma spp.	Water-plantain	Kobresia capillifolia	= Carex capillifolia
Andropogon gerardii	Big blue stem	Kobresia humilis	= Carex alatauensis
Andropogon lateralis	Beard grass, bluestem grass, broomsedge	Kobresia pygmaea	= Carex parvula
Arnica montana	Wolf's bane, leopard's bane, mountain tobacco, m. arnica	Kobresia tibetica	= Carex tibetikobresia
Artemisia frigida	Silky wormwood	Leymus chinensis	Chinese ryegrass
Artemisia ordosica		Leymus triticoides	Creeping wildry
Artemisia sieberi		Lolium multiflorum	Italian ryegrass
Artemisia tridentata	Sagebrush	Lolium perenne	Perennial ryegrass, English ryegrass
Hordeum leporinum	Barley-grass	Lollium spp.	Ryegrass
Atriplex lentiformis	Quail bush, big saltbush	Lotus corniculatus	Birdsfoot trefoil
Avena barbata	Slender wild oat, bearded oat	Medicago polymorpha	California burclover, toothed bur clover, or toothed medick
Avena strigosa, Avena fatua	Black oats	Medicago sativa	Alfalfa
Axonopus affinis	Common carpetgrass	Medicago spp.	Medick, burclover
Bassia dasyphylla	Shaggy-Leaved Bassia	Megathyrsus maximus	Guinea grass cv. ‘Mombaça’
Bouteloua gracilis	Blue grama	Nardus stricta	Matgrass
Brachiaria brizantha	Palisade grass (‘Marandú’)	Ornithopus compressus	Yellow bird's-foot
Bromopsis inermis	Smooth brome	Panicum antidotale	Blue panic, giant panic-grass
Bromus spp.	Brome	Pascopyrum smithii	Wheatgrass
Calamagrostis brachytricha	Feather reed grass, foxtail grass, diamond grass	Paspalum spp.	Bahiagrass, crowngrass or dallis grass
Calamagrostis spp.	Tussock grasses	Paspalum dilatatum	Dallis grass
Carex atrofusca	Dark brown sedge or scorched alpine sedge	Paspalum notatum	Bahiagrass
Carex moorcroftii		Paspalum piptochaetium
Carex sempervirens	Evergreen sedge	Paspalum vaginatum	Paspalum, seashore paspalum
Carex spp.	Sedge grass	Phalaris arundinacea	Reed canary grass
Carissa edulis	Climbing num-num, simple-spined num-num	Phleum pratense	Timothy grass, cat’s tail
Carya spp.	Hickory	Pinus koraiensis	Korean pine
Celtis sp.	Hackberry	Pinus pinaster	Maritime pine, cluster pine
Cerastium spp.	Mouse-ear chickweed	Pinus pinea	Stone pine, Roman pine, parasol pine, umbrella pine
Chrysothamnus nauseosus	Chamisa, rubber rabbitbrush, and gray rabbitbrush	Pinus ponderosa	Ponderosa pine
Cirsium spp.	Thistle	Plantago lanceolata	Buckhorn plantain
Crotalaria juncea	Sunn hemp	Poa angustifolia	Narrow-leaved meadow grass
Cynodon dactylon	Bermudagrass	Poa pratensis	Kentucky bluegrass
Cynodon dactylon × C. transvaalensis	Hybrid Bermudagrass	Poa spp.	Meadow-grass, bluegrass, tussock and speargrass
Cynosurus cristatus	Crested dogtail grass	Polylepis spp.	Tabaquillo
Cytisus spp.	Broom	Populus euphratica	Euphrates poplar
Dactylis glomerata	Cat grass, cocksfoot	Populus spp.	Poplar Tree
Deschampsia cespitosa	Turfed hair grass	Prosopis spp.	Mesquite
Dichanthelium spp.	Witch grass	Quercus faginea	Portuguese oak
Dodonea angustifolia	Sand olive	Quercus ilex	Holm
Elymus nutans		Quercus rotundifolia	Holm
Elymus smithii	Wildrye, wheatgrass, squirreltail	Quercus spp.	Oak trees
Erica spp.	Heaths	Trifolium repens	White clover
Fagus sylvatica	Beech	Trifolium resupinatum	Persian clover
Festuca arundinacea	Tall fescue grass	Trifolium subterraneum	Subterranean clover
Festuca glauca	Blue fescue		Short bunchgrass
Festuca nigrescens	Chewing’s fescue