1Dept. of Land, Air and Water Resources 2Dept.of Agronomy & Range Science University of California, Davis, CA 95616 Submitted to: Global Change Biology July 1996

Abstract

Introduction

Evapotranspiration is directly affected by climate factors affecting the energy budget, including net solar radiation, temperature and precipitation. Higher evapotranspiration demand is predicted under higher temperature. In California, daily reference evapotransipiration (ET_o) is one of the most useful guides for scheduling crop irrigation. The actual ET is estimated from ET_o and crop coefficients K_c, using an equation that takes the form: ET=K_c* ET_o. This type of linear scaling equation combined with the summed acreage provides a good indicator of crop water demand. It also provides information on the maximum potential yield if the water supply is sufficiently near ET. Therefore, ET_o is widely used by California water districts to allocate agricultural water to growers (reference from Hanson and Dave G? in LAWR).

Alfalfa is one of the most important crops in California despite having a high water demand because it fixes nitrogen, thereby maintaining forage quality for livestock production. California produces 8.4% of the United States alfalfa hay production on only 4% of the nation’s alfalfa acreage. The yield is primarily limited by water availability during the growing season in California. Alfalfa is a perennial crop and has a relatively complex growth pattern since the site yield accumulates over several cutting and regrowth cycles. The yield is composed of both leaf and stem tissues and the proportion of each determines the forage quality. The size of the crown and taproot and their reserves of carbon and nitrogen substrates influence the number and vigor of the regenerated stems after each cutting. Other factors such as cultivar characteristics also influence alfalfa yield. Most of the alfalfa (80%) was grown in the San Joaquin Valley and the low desert regions because of warm weather and a long growing season. However alfalfa is grown in four zones (Intermountain, Sacramento Valley, San Joaquin Valley, and Low desert zones) grouped according to the climate and geographic conditions (Teuber 1984).

In general, the Intermountain zone including Tule Lake, Butte Valley, and Scott Valley produces about 11% of the alfalfa in California. The first cutting usually takes place in mid-to-late May and is followed by 2 or 3 more cuttings per year producing an annual yield of 1836 - 2570 kg ha^-1 and requiring approximately 76 cm water per growing season. The Sacramento Valley zone, including Sacramento and Yolo Counties, produce 8% of the state’s alfalfa. The first cutting occurs in April with a 30 day cutting interval initially and 28 days in summer. The annual yield is about 2570 kg ha^-1 with 6-7 cuttings. The San Joaquin Valley zone, which includes Five Points and the Kearny Agricultural Center and provides 60-80% of California alfalfa. Established stands are cut for the first time in early April and harvested through October. Growers typically get 8 cuttings per year with an average yield of 3305 kg ha^-1. The sensitivity to water limitation is illustrated by the yield decline to 2754 kg ha^-1 during the 1991 drought. The Low Desert zone, represented by El Centro, produces 20% of the California alfalfa crop. The climate in this area is warm with little precipitation (about 7.5 cm) a year. The annual yield is about 10 tons per acre with about 9-10 cuttings. The first cutting at El Centro starts in January or February, with cutting cycle of 28 days in summer and 6-8 weeks in winter.

If the predicted climate changes are realized, then California alfalfa production will be seriously affected and alternative agricultural practices will need to be developed in order to sustain current levels of productivity. This study assesses the potential changes in temperatures, precipitation and ETo under elevated CO₂ scenarios, in order to evaluate the potential impact on alfalfa crop production, and to anticipate needs for agricultural water reallocations and possible changes in alfalfa crop distributions in California.

Materials and Methods

The 20 - 30 year (depending on the length of weather station history) long-term monthly mean values of 114 weather stations in the Central and Imperial Valleys were summarized from the CIMIS database maintained by the California Department of Water Resources. Data from months having more than five missing days were not included for the calculation of monthly averages, which were used as input data for the SIMETO program. One hundred and fourteen weather stations (Figure 1) were used for the calculations of temperatures and precipitation while only 80 weather stations had observed ETo data for simulation estimations and comparisons. Statistical relationships were developed for interpolating GCM daily temperatures and monthly precipitation between weather stations because the direct outputs from GCM for each grid cell is too coarse for California crop assessment (California regions were covered in only 12 grid points from GCM). R² values greater than 0.98 were obtained for the maximum and minimum temperature adjustment relationships for all months in this study. The spatial interpolation relationship for the precipitation adjustment was statistically significant but with lower R² (0.6-0.75) values for the months of January and Febuary adjustments, and higher R² (0.8 - 0.95) values were abtained for other months in precipitation adjustment interpolations.

Daily ETo was estimated using nine different methods (Table 1, Snyder and Pruitt 1994) for the 1xCO₂ condition in the simulation model. Then the ETo was averaged from all nine predictions under 1xCO₂ condition for the assessment. However, because the weather input information was not sufficient to estimate ETo for all nine methods under 2xCO₂ conditions in the weather model (Table 1), ETo was simulated from only three methods. To be consistent in the assessment, then the average ETo for the 2xCO₂ conditions was estimated through a regression relationship that was established by ETo under the 1xCO₂ conditions. In other words, a relationship was constructed from the average ETo derived from the nine methods used in 1xCO₂ condition and the average ETo constructed from the three methods that were available to calculate ETo under 2xCO₂ condition. Moreover, the ETo under 2xCO₂ condition was also calibrated using a calibration factor between simulated ETo under current CO₂ scenario and the actual ETo observed from each weather station. Then the calibrated ETo in 2xCO₂ was compared to the actual ETo in the assessment.

Six locations were selected from different climatic zones for comparisons: Tule Lake and Gerber Dryland (Intermountain zone), Davis (Sacramento Valley), Five Points (San Joaquin Valley), Santa Barbara (Coastal area), and El Centro (Low derset zone). Because little alfalfa grows in the Coastal zone such as at Santa Barbara, and Gerber Dryland had incomplete weather data, only four locations were compared in the ALFALFA simulation. Current CO₂ and 2xCO₂ scenarios, and two water management practices were used as inputs to the model for this study. The current water practices for each location were obtained from the University of California Cooperative Extension specialists and farmer advisers (Table 2). Production simulations assumed that water supplies were not limited. The alternative water management was determined subjectively based on the weather conditions generated by SIMETO (Table 3). The harvest scheduling under 2xCO₂ conditions was adjusted according to the temperature requirements for initiation of spring growth, and the interval required between cuttings and the final cutting date. The dates for initiation of spring growth and for the last harvest were assumed to occur at the same mean minimum temperatures as used under current weather conditions. As a result, the harvest interval was adjusted to produce approximately the same number of days between cuttings as used under current weather conditions. Using the long-term monthly summary data for minimum and maximum temperatures and precipitation, thirty-years of weather data were simulated. Alfalfa yields were estimated using these simulated thirty-years of weather data as model inputs under each defined scenario.

Several physiological functions in the ALFALFA model were modified to account for the direct effect of CO₂ on alfalfa yield such as intercellular CO₂ concentration and stomatal opening. The light response function of photosynthesis (Figure 2) was modified using coefficients derived from clover grown under 2xCO₂ conditions (Ryle et al. 1992). The effect of relative water content on stomatal opening was reduced for 2xCO₂ conditions (Grant et al. 1995, Kimball et al. 1995).

Results

1. California Weather and Reference Evapotranspiration

In the selected alfalfa regions, Intermountain zone (Tule Lake, northern California), had the largest increases in maximum and minimum temperatures. Five Points in the central San Joaquin Valley had the second largest increase in minimum temperatures (Table 4). Precipitation increased slightly for Five Points and Davis in the middle of the Central Valley, but did not increase for any other regions under 2xCO₂ conditions (Table 4). However, there was not sufficient input data for Tule Lake and Gerber Dryland to simulate precipitation under 2xCO₂ scenario. The expected precipitation changes are to increase northward and to decrease southward. Monthly patterns of maximum and minimum temperatures (Figure 3) were similiar at all four locations under both 1xCO₂ and 2xCO₂ conditions. Both daily mean maximum and minimum temperatures increased more in the summer and autumn than was observed for the winter or spring seasons at the four locations. Precipitation patterns were higher in the opposite seasons, which corresponds with the current rainy season (Table 5).

The daily mean ETo was estimated to increase by 0.59 mm. The increase in ETo under 2xCO2 conditions was larger in summer and autumn than in winter and spring (Table 5). Figure 4 shows the monthly ETo at the five alfalfa regions. The increase in ETo was highest in the months of June, July, August and September (Figure 4). The increase in ETo in the Coastal Range was smaller relative to the sites in the Central Valley. There was no significant difference between ETo values among the regions within the Central Valley. Moreover, although the increase in ETo was smaller in the winter and spring seasons, the percent change was higher in these two seasons (Table 5).

The seasonal variations in temperature, precipitation, and ETo under both climate scenarios were larger in spring and winter than in summer and autumn. ETo of 1xCO₂ had larger variations than ETo of 2XCO₂ climate. However, the magnitude of variation was larger under 2xCO2 conditions for temperature and precipitation. Moreover, precipitation exhibited the greatest range among all variables under both current 1xCO₂ and 2xCO2 conditions.

Considering the water balance, the long term water supply must equal the water demand. Therefore, according to the water budget, precipitation + water project (irrigation) supply = ET + municipal needs + surface run off (this is minimal in the alfalfa growing areas). Annual mean precipitation will increase under 2xCO₂ conditions by 1.42 mm monthly while ET will increase by 17.7 mm monthly. Assuming other factors remain the same, the discrepency between increased precipitation and ET will be -0.7 mm monthly. California has 123,828 km² agricultural land and alfalfa is grown on 3.3% (4000 km²). Monthly water supply must be balanced by 2*10⁹ m³ of irrigation water for agricultural land to maintain the current cropping systems and production level. Therefore, the total extra water needed annually for agriculture would be 2.4*10¹⁰ m³ which is equivalent to the ?? percent of Folsom ? reservoir water. For the alfalfa agricultural region of California, 6.6*10⁷ m³ in monthly water supply will be needed to maintain the current alfalfa production during the growing period. This demand will be greatest in the summer and autumn months because there is little or no precipitation and the largest ET occurs during these months (Table 5).
 

2. California Alfalfa

Figure 5 shows that yields increased in the spring and summer at each location, then decreased in the late summer and in the autumn periods. Alfalfa at El Centro has the longest growing period, Five Points is the second longest season, followed by Davis, and Tule Lake with the shortest growing period. These yield responses follow the latitudinal gradient in growing season length. In El Centro, yields of the first two harvests slightly increased but then decreased for harvests over the rest of the season under doubled CO₂ conditions. These results indicate that the present cultivar will not adapt to an elevated temperature environment. The direct effect of temperature on yield was most apparent at Tule Lake under current conditions while at Five Points and El Centro, the yield per cut was more closely related to cumulative solar radiation.

Discussion

ETo is affected primarily by temperature. After changing the simulation data input and running SIMETO without precipitation, the simulations showed that total precipitation has little direct effect on ETo under the California climate conditions, because precipitation is primarily restricted to winter months. Winter precipitation does not affect ETo because of the high humidity, lower temperatures, and lower net radiation (Rn) due to overcast skys during winter months (Rn= ET + sensible heat (H) + storage (S)). With the higher estimated ETo under 2xCO₂ climate, current cropping systems may change because of limited water allocations, especially in southern California.

Evapotranspiration during winter consists of evaporation from wet soils or water bodies following precipitation events and from transpiration from agricultural and native vegetation, especially winter grasses. Grasslands and rangelands are found throughout California in areas with limited annual precipitation, generally less than 250 mm. During the summer, high evapotranspiration is associated mostly with irrigated agriculture. Therefore, increased temperatures would result in enhancement of winter evapotranspiration by native vegetation and summer evapotranspiration by agriculture which would lead to greater total water demand in California.

With predicted climate changes, including larger percent increases in evapotranspiration and smaller increases in precipitation-- clearly increased demand is not proportional to the increased supply. These predictions will result in reducing the supply of soil water if water allocation and the frequency of irrigation does not increase to meet the deficit. Crop growth is predicted to be negatively affected through the biological and physiological responses to more limited water supply.

Although increasing temperatures increase evapotranspiration demand in irrigated agriculture, the increased ambient CO₂ might also increase photosysthetic rate in crops and partially compensate additional water demand (Ryle et al. 1992, Grant et al. 1995). This affect minimizes the otherwise larger decline in net photosysthesis for alfalfa crop under 2xCO₂ scenario. This CO₂ fertilization effect has been taken into consideration in the ALFALFA model for 2xCO₂ condition. Moreover, it is possible that the redistribution of precipitation may benefit some agricultural regions, since more precipitation is expected in the Central Valley. The simulated alfalfa model illustrated that yield increased if water supplies were sufficient in northern California, while yield decreased even under the alternate water mangement strategies in southern California. These results suggest that to maintain productivity, the physiology of the cultivar will be as important a factor as resource allocation under a 2xCO₂ altered climate.

Given the amount of water required (Table 5) to supplement the alfalfa production system, it is likely to increase demand for groundwater pumping. If groundwater is to provide the main water source to balance the irrigation demand, then it is likely to lead to rapid groundwater depletion with subsequent overdrafting and land subsidence. Early snowmelt will increase surface runoff in the spring relative to current snowmelt. Increasing winter and spring temperatures may well decrease the frequency and intensity of the snowpack in the winter as well. These changes suggest the potential for decreasing the summer surface water supply. If more dams or reservoirs were built to retain more of California's water resources it might be to provide more summer irrigation water. However, habitat loss following diversion of additional surface waters may produce sufficiently greater stresses on the production and diversity of wildland species that it could severely impact the natural environment and cause further environmental degredation. Environmental and urban competition for water resources may prevent substantial increases in agricultural water allocations. Therefore, water diversion may not be a solution to predicted water shortages for sustainable agriculture. Cropping systems will have to change in order to balance water supply to meet projected water shortages,. Alfalfa is a high water consuming crop and its distribution is likely to shift northward to maximize the yield per water unit.

Minimum daily temperatures are more critical to alfalfa growth during spring and autumn in the colder areas while solar radiation is the most important factor determining alfalfa growth in warmer areas because high temperatures are less limiting than low temperatures. The expected increases in temperature enables alfalfa to start growth earlier in the year but to demand more water in the summer. The yield differences predicted between colder and warmer regions may be due to differential sensitivity to temperature and solar radiation during the growing season. The yield of the last few harvests, shown in Figure 5, were lower than expected from the effects of temperature alone, indicating the impact of day length and total daily net radiation. This pattern was observed in all four alfalfa regions. At Five Points and Davis, under 2xCO₂ conditions, the yield declines in autumn harvests were primarily attributable to the shorthened daylength and the resulting reduction in solar radiation and secondarily, to the redistribution of photosysthate to the roots. In Tule Lake, the duration of the period of optimal temperatures for alfalfa growth was increased under 2xCO₂ conditions, which combined with longer autumn day length, increased the total yearly yield at this location. The most noticeable yield reduction attributable to high temperatures occurred at El Centro. At this location, yields are reduced during the summer months as well as in the fall. This may explain why yields decreased in the warmer areas under 2xCO₂ scenarios with the current management. With increased temperature and precipitation in the spring months, more early alfalfa growth is likely to occur and the harvest of the first alfalfa cut should occur earlier than under the current condition in order to make full use of the available resources. Therefore, the yield increased most significantly in the coldest areas when alternative cultural practices are used. Figure 5 also showed that the yield declined at the last growth harvest, partly attributable to an underestimation from alfalfa model (Denison and Lommis 1991).

The percent increase in temperature and ETo is not accompanied by a proportional percent increase in the precipitation water supply. Therefore, water resoruces will be limited to meet the agricultural demand estimated for alfalfa production. If these climate predictions are robust, we must be prepared to change the cropping system to meet the water resource supply. Based on model predictions for California agriculture as a whole, more emphasis should be placed on winter crops that would best utilize the available water resources while minimizing summer crops, at least in southern California. Such changes are necessary because of expected water shortages and high evapotranspiration rates. More dryland crops may be needed and the replacment of rice paddys in the Sacramento Valley to to other crops reduce agricultural water comsumption. Such crop shifts were observed during the recent drought (1987-1993) when water allocations were severely reduced and more land was used for cotton production or left fallow. It seems clear that if climate scenarios are correctly predicted, anticipation of northward shifts in cropping systems should be considered to maximize water resource use. Optimal growing temperatures for a given crop will be shifted to more northern latitudes, a growth effect enhanced by increased daylengths. These advantages are offset by greater seasonal variability. However, if the California water supply were sufficiently increased, southern California could increase alfalfa production to a double cropping system.
 

Acknowledgments

References

Barry TA, Geng S (1993) The effects of climate change on United States rice yields and California wheat yields. Climate Change and Agriculture: Analysis of Potential International Impacts, ASA Special Publication Number 59.

Denison F, Loomis RS (1991) An integative physiological model of alfalfa growth and development. University of California, Division of Agriculture and Natural Resources Publication 1926. Oakland, California.

Doorenbos J, Pruitt WO (1977) Crop water requirements. FAO Irrigation and Drainage Paper 24, United nation Food and Agriculture Organization, Rome, Italy.

Eastering WE, Crosson PR, Roserberg NJ, McKenney MS, Katz LA, Lemon KM (1993) Paper2. Agricultural impacts and responses to climate change in the Mosouri-Iowa-Nebraska-Kansas(MINK) region. Climate Change, 24, 23-61.

Geng S, Auburn J, Brandstetter E, Li B (1988) A program to simulate meteorological variables: Documentation for SIMMETEO. Agronomy Progress Report No. 204, University of California, Cooperative Extension, Davis, California.

Grant BA, Garcia RL, Pinter PJ, Hunsaker D, Wall GW, Kimball BA, LaMore RL (1995) Interaction between atmospheric CO₂ concentration and water deficit on gas exchange and crop growth: tesing of ecosys with data from the Free Air CO₂ Enrichment (FACE) experiment. Global Change Biology, 1, 443-454.

Kimball BA, Pinter Jr. PL, Garcia RI, LaMorte RT, Wall GW, Hunsaker DJ, Wechsung G, Wechsung F, Kartschall T (1995) Productivity and water use of wheat under free-air CO₂ enrichment. Global Change Biology, 1, 429-442.

Idso SB, Kimbal BA, Anderson MG, Mauney JR (1987) Effects of atmospheric CO₂ enrichment on plant growth: the interactive role of air temperature. Agriculture, Ecosystem and Environment, 20, 1-10.

Kellomaki S (1995) Computations on the influence of changing climate on the soil moisture and productivity in Scots pine stands in Southern and Northern Finland. Climatic Change, 29, 35-51.

Mitchell JFB, Manabe S, Meleshko V, Tokioka T (1990) Equilibrium climate change and its implications for the future. In: Climate Change (eds Houghton JT, Jenkins GJ, Ephraums JJ), pp. 131-175. Cambridge University Press, Cambridge.

Mount TD (1994) Climate change and agriculture: a perspective on priorities for economic policy. Climate Change, 27, 121-138.

Orr DW (1993) Agriculture and global warming. Agriculture, Ecosystem and Environment, 46, 81-88.

Rawlins SL (1991) Global environmental change and agriculture. Journal of Production Agriculture, 4, 291-293.

Rosenberg NJ (1988) Global climate change holds problems and uncertainties for agriculture. In: US Agriculture in a global setting, An Agenda for the future (ed Tutwiler MA), Resources for the future, Washington D.C.

Ryle GJA, Woledge J, Tewson V, Powell CE (1992) Influence of elevated CO₂ and temperature on the photosynthesis and respiration of white clover dependent on N₂ fixation. Annuals of Botany, 70, 213-220.

Schlesinger ME, Zhao Z (1989) Seasonal climate changes induced by double CO₂ as simulated by the OSU Atmospheric GCM/Mixed-Layer ocean model. J. Climate 2, 459-495.

Smith M (1992) Report on the expert consultation on revision of FAO methodologies for crop water requirements. United Nations - Food and Agriculture Organization, Rome, Italy.

Snyder RL, Pruitt WO (1994) User’s guide and program documentation. Department of Land, Air and Water Resources, University of California, Davis.

Tsonis AA (1991) Sensitivity of the global climate systems to initial conditions. EOS 72, 313.