Quantifying Plant Water Use in the Native Coteau Rangelands of North Dakota

Results for 2001



Plant water use is an important part of the ecosystem research we are conducting at the CGREC. To successfully simulate patterns of plant water use under a rangeland situation, we first need a good estimate to some crucial parameters (both of plant and soils) governing various flows of water in the soil-plant system. This year, we measured soil water characteristics, plant stomatal conductance and photosynthetic water use efficiency, and leaf area index.

 

Soil water profile anatomy

 

First, we will see how soil water content changes on different range sites at different times of the year 2001 (Figure3). In early and mid May, all four sites have a wet layer at a depth 0.25 to 1.75 ft. The wet layer in sites 11SX and 11D shifted upward and peaked at 1.75 ft. Melted snow should have some contribution to the formation of this springtime wet soil layer. A major rainfall on 5-6 May 2001 caused an increase in soil water in the shallow layers of all four sites. Sites 5D and 5C had a slight increase in water content which decreased gradually. The relatively low water content in the surface layer of these sites could be due to evapotranspiration (loss of water through plants and soil surface especially during the daytime period) or due a rapid drainage of moisture to deeper layers. On 15 May a pulse of water content occurred at layers 0.75 ft to 2.25 ft in site 5C which might be caused by other sources (i.e., some animal interference) because there was no rainfall between 8 May and 15 May. The different patterns of response between sites 11SX and 11D can be explained in terms of different surface cover. There is a very thick litter layer in site 11SX that may retard rainwater infiltration into deeper layers but also prevent rapid water loss from the surface layer (0.25 ft) following the rainfall event. In contrast, the surface-layer water content in 11D recorded on 8 May (3 days after the rainfall event) is much lower than in 11SX. From 8 to 15 May, the first two layers of site 11D (0.25 ft and 0.75 ft) had a considerable drop in water content, while during this time period the water increase in deeper layers (1.25 ft and down) was barely noticeable or not changed. This implies a contribution from evapotranspiration. Considering the relatively sparse vegetation cover at this site, it follows that soil evaporation in this extremely grazed site is important.

 

Soil water content profile in the fall 2001, as indicated by observations made on 18 September and 1 October is drastically different from the spring pattern: the wet plateau which prevailed in the spring disappeared and water content ropped uniformly from about 2.75 ft to 0.25 ft, except for the site 11SX where a modest plateau was still maintained at depths 0.75 ft to 2.25 ft. This drop of soil water was brought about mainly by (1) plant water use during the major part of growing season (from early May to early Oct.) and (2) soil evaporation caused by the very hot summer plus a scarcity of rainfall from summer to early fall season. The moderate drop of soil moisture as observed in site 11SX is apparently the result of low evapotranspiration in this site, because, in addition to the low soil surface evaporation (due to the existence of surface litter layer), plant transpiration should be very low in this site (about 99.74% of leaf area index was attributed to Kentucky bluegrass in a similar non-grazing exclosure (Fig.7D) and this plant is physiologically inactive with a low stomatal conductance in this very dry season (Figure 5)).

 

Another interesting point to note is that, in three of the four sites (5C, 11SX, 11D), the differences in soil water content from 2.75 ft down to 8.5 ft between May and October measurements were small, while in site 5D the stable water content occurred only at 4.5 ft and deeper. Further analysis is required to see if the soil water actually remained unchanged from early May to early October in this year. In addition to soil water status, this site (5D) also has a low plant cover and physiological activity (See Figure 5). This background information should be accounted for in the next year’s drought experiment.

 

Soil physical properties and water characteristic curves


Table 1 shows the measured soil physical properties for samples from three sites (6X, 5D, 11D). The percentage of soil particles are classified as sand, silt and clay. The definition of the three size-classes (one micron equals to one millionth of a meter) are shown in Table 1. Bulk density (BD) is defined as the mass of dry soil contained by a unit of fresh soil volume. The bulk density for sandy soils (typical value 1.6) is higher than for clayey soils (typical value 1.3). “Ksat” is a measure of how fast the water moisture is potentially able to move when soils are fully saturated with water. The top 0-6 in. layer differed from the lower layer in % organic matter, bulk density and in Ksat, but the trend is not clear in the size-class percent data. The difference most likely is related to the high root activity in the top 0-6 in. layer. At site 6X, percent organic matter dropped from 6.9% to 3.2% from 0-6 in. layer to 6-12 in. layer. Then it continued to drop with the increase of depth. The bulk density for 0-6 in. layer was less than 1 g cm-3, while for the deeper layers, it was higher. The Ksat for the first layer was about 100-1000 times that of the 30-36 in. layer. The highest bulk density was obtained at site 6X (24-30 in. layer, BD 1.738 g cm-3). This value is an unusually high value as seen in this grazing intensity site. The measured bulk density for sample “6X12-18” is 1.332 g cm-3, however, in the laboratory when this sample was tested for water characteristic curve, the value 1.322 g cm-3 was used instead. So, we used the latter value (1.322 g cm-3) in the curve-fitting procedure (assuming the effect of this change is minor).

 

The soil water characteristic curve, also called soil water retention curve, is the relationship between soil water potential and soil water content (Hillel, 1982). The basic structure of soils as porous media are the array of various sized particles forming the basic matrix in which plant roots and other soil resources (water, solutes, and organic matter) are imbedded. Soil water potential is a measure of energy status of water molecules held in the soil particle interstices. It determines how easily it is for water to move from one location to another in the soil profile, as well as how easily soil water can be extracted by plant roots. Soil water content is the quantity of water (in mass or volume basis) held by the soil porous media. From these definitions, the importance of the soil water characteristic curve to plant water use is apparent.

 

We fitted the following two equations, describing soil water characteristic curves, to our measured data using the non-linear least squares curve-fitting method. The first equation was proposed by Campbell (1974) and is used in the ecosystem model by Thornley (1998). In the following discussion, it is referred to as “the Campbell equation.”

y = ymax (qmax/q)q [1]

 

where y is soil water potential; q is soil water content; ymax is the maximum soil water potential when soil is fully saturated with water (qmax); q is an empirical parameter to be determined. The second equation is

 

 Se = [1+(y/ymax)n]-m [2]

 

where y is soil water potential, Se is the effective saturation and is defined as Se = (q-qr)/(qmax-qr), qr is the residual water content (the amount of soil moisture that is so tightly held by the very fine particles that even oven-drying (at about 100 °C) can not remove this amount of moisture), ymax and qmax are maximum water potential and water content, respectively, and n, and m are empirical parameters to be determined. This equation was proposed by Van Genuchten (1980) and is discussed by Van Genuchten and Nielsen (1985). In the following text, we will refer it as “the Van Genuchten and Nielsen equation”, or sometimes abbreviated as “GN”.

 

As shown in the root mean square deviation (RMSD) columns in Table 2, both the Campbell equation and the GN equation fit the measured data reasonably well.. Generally, the curve-fitting procedure gives out reasonable parameter estimates. However, some uncertainties exist, especially in qmax. In the modeling efforts of Thornley (1998), this term was treated as “field capacity”, it may be better to define it as the “maximum water content” which can be higher than field capacity (Campbell, 1974).

Now let’s choose the Campbell equation to extrapolate the measured soil water potential-water content relationship to drier regions and to see its implications to plant water use. The comparison will be made between samples from 0-6 in. (with a high root activity and organic matter) and from 30-36 in. (with low root activity and low organic matter). The results are shown in Figure 4. The horizontal axis is the volumetric water content which is the product of mass-based percent water content and bulk density. The vertical axis is water potential in Joule per kg of water. In the soil matrix the maximum water potential is zero (soil fully soaked with water). As soil becomes drier and drier, water potential becomes more negative. Usually the values of –0.33 bar (i.e., –33 J kg-1) and –15 bar (i.e., –1500 J kg -1) are considered the upper and lower limits, respectively, of soil water potential, within which soil water is usable by many plants. This range is shown as vertical bars in Figure 4. (for some drought hardy plant species the lower range can be much lower than –1500 J kg-1). It can be seen that an interesting difference exists between the curves of the top layer (0-6 in.) and deeper (30-36 in.) in each of the three sites. Beginning from the zero potential, when soil water potential begins to drop (for example, due to plant water use), the top layer’s water content decreases faster than the deeper layer in the wet region of the curves. Then in the relatively drier region of the curve, say beginning from 30% water content, the speed of the decrease in water content for the top layer (as water potential continues dropping) reduces, or even could be slower than that of the deeper layer. This is to say that, when the soils are relatively wet, the top layer’s capacity to provide water to plants is potentially lower than the deeper soil layer. When the soils become dry, the top layer’s capacity of releasing water for plant use tends to increase. However, as soil dries, it is more difficult for plants to extract the water held by the fine soil particles. Despite the advantage of the deeper soil layer in water characteristic, root systems of many range plants usually can not reach deep enough to make use if this advantage. Water flow from the deeper layer to the top layer may be very slow as indicated by the very low hydraulic conductivity of this deeper soil layer (Table 2). As a result, the top layer of the soil (0-6 in.), with its poor water holding capacity, may frequently have a problem of water deficit under a high evapotranspiration demand. It is possible that the particular shape of water characteristic curves of the top soil layer is attributable to the altered pore-size distribution caused by the high organic matter content.

   

Plant water use efficiency and stomatal conductance


Water use efficiency (WUE) is defined as photosynthetic rate divided by transpiration rate for leaves of the plant under consideration. Photosynthetic rate is the amount of CO2 fixed by the plant leaves per unit leaf area per unit of time. This parameter is important because production agriculture, no matter what is concerned (grasses, crops or animals produced), relies ultimately on the ability of green plants to use atmospheric CO2 as the main raw material to build biological structure. Photosynthetic rate is closely related to the speed of this building process. Transpiration rate is the amount of water lost from plant leaves to the free air per unit leaf area per unit time. Stomatal conductance is a measure of how easily water molecules can escape from inside of plant leaves through the tiny pores (stomata) on the leaf surface to the free air. It is highly positively correlated with transpiration rate (the unit “mol” means “mole”: for example, one mole of water equals 6.022 ´1023 H2O molecules).

Beginning this year, we are measuring the above mentioned gas exchange parameters for selected range plants in the field. The purpose is to obtain some quantitative understanding on how these parameters are related to factors like temperature, soil water content, etc. This year, we measured for 15 days at different times of the year (spring, summer and fall) on about 16 species, but the majority work was focused on Kentucky bluegrass. The data has not been fully analyzed yet, but we have highlighted some interesting results in Figure 5. These are the results from 24-25 September (see Table 3 for average radiation and temperatures) when very low soil water content was observed on the silty sites (Figure 3). In this drought situation, the majority of the Kentucky bluegrass leaves became folded together (KTKFOLD) with a low stomatal conductance and an extremely low photosynthetic rate. The resultant water use efficiency, as indicated in Figure 5 is very low. A minor portion of the Kentucky bluegrass leaves (accounting for about 10% of the total Kentucky bluegrass leaf population), however, did not fold, and were leaves produced by fall season regrowth. The photosynthetic rate and stomatal conductance of the unfolded leaves tend to be higher than that of the folded leaves and somewhat comparable with the then actively growing western wheatgrass (Figure 5), although the latter (western wheatgrass) tended to have a higher stomatal conductance. Because of the very low soil water potential (Figure 3), the high physiological vigor of western wheatgrass suggests that its roots may reach deeper in the soil layer, compared with Kentucky bluegrass. One interesting question to ask about Kentucky bluegrass is how did the regrowth of this plant absorb water from the extremely dry top soil layer given the usually shallow rooting habit of this grass species?

 

Figure 6 highlights the results measured in the CRP field for two grasses: tall wheatgrass (Agropyron elongatum) and intermediate wheatgrass (Agropyron intermedium) and two legumes alfalfa (Medicago sativa) and yellow sweetclover (Melilotus officinalis). These four species were planted 13 years ago when cropland was converted to the CRP field. What is the possible long-term dynamics of this CRP plant mixture, in terms of the competition and co-existence between the grass and legume plants? Research in southern England by Thornley et al. (1995) indicated that physiological aspects like leaf photosynthetic capacity and root nitrogen uptake kinetics have a major influence on the long-term stability and co-existence of the grass-legume mixture. We will test this possibility in the CRP field of the North Dakota grasslands. We hope the field data could reveal some possible differences in photosynthetic water use of these four species. As indicated in Figure 6A, the differences between species were not clear on 22 Jun. (see Table 3 for radiation and temperature), when all four species were expanding their leaves. However, in photosynthesis, tall wheatgrass tended to have a higher photosynthetic rate than yellow sweetclover (Figure 6). On 25 August (see Table 3 for radiation and temperature), tall wheatgrass’ photosynthetic rate dropped, as did its water use efficiency, while alfalfa still kept a considerable photosynthetic and water use efficiency (Figure 6). On this date, it’s difficult to find healthy leaves of yellow sweetclover and intermediate wheatgrass to measure. From these preliminary data, it seems that alfalfa is photosynthetically competitive in both the summer and fall season in this CRP field, while the other legume, yellow sweetclover, is not. The two grasses species’ photosynthetic activity peaked in the early summer (for depth 0-12 in. on this CRP site, soil bulk density was measured as 1.58 g cm-3, volumetric water content was 0.13; for depth 12-24 in., bulk density was 1.53 g cm-3 and water content was 0.27). Further analysis of the data is required to confirm these overall patterns in water use among the four CRP species and further data is required to see how these trends change over time.

Leaf area index for selected plant stands

 

Leaf area index, the amount of green leaf per unit of ground area, is another very important physiological characteristic that strongly influences plant water use and dry matter production. In the mixed-grass prairie and especially under grazing pressure, plant coverage can be very low. It is hard to use some commercially available equipment to automatically measure leaf area index. So, beginning this year, we will manually measure this parameter for selected range sites (see “Methods of Field Measurements” for detailed measurement procedure). Figure 7 is a complete summary of this year's measurement (see Table 4 for list of plant name codes used in Figure 7). We need to note here that the “leaf area” reported here is the “green” or “live” leaf area that existed in the particular plant communities. The dead leaves were not accounted for in this measurement. Due to the limit of manpower, we did not replicate measurement for each of the plant stands (eight in total). But the general trend revealed by this estimation seems reasonable: (1) The results in the CRP field (Figure 7A) are consistent with the physiological measurement as shown in Figure 6. (2) For three Kentucky bluegrass communities (Figure 7B-D), we can see the dominance of Kentucky bluegrass increases as grazing pressure decreases (in the non-grazing exclosure, almost a pure Kentucky monoculture was observed throughout the whole growing season). Also, it’s clear that plant species composition is more diverse under increased grazing pressure. This is especially so in the extremely grazed pasture (11C) (Figure 7B). Also shown in Figure 7B is the gradual decrease of the percentage of the “miscgra” (miscellaneous grasses) from early summer to fall seasons, suggesting this category of grasses seems to be favored by cattle over Kentucky bluegrass. In Figure 7E-H, it can be seen that, as fall season approached, leaf area index of stiff goldenrod stayed at a higher value than that of buckbrush.

 

These trends will be useful for calculating seasonal water consumption by these different types of plant stands. Specifically, as has been stated briefly in the “Methods of Field Measurements” section, these measured leaf area indices will be useful for estimating the canopy conductance (average conductance for water vapor for the whole canopy composed by a mixed plant species with different stomatal conductance each), which is used for the Penman-Monteith equation to calculate evapotranspiration (Thornley, 1998).

 

Concluding remarks

 

At this first stage of research we do not have information related directly to the management of rangeland. However, the field measurements as explained in the above paragraphs have revealed some interesting points that can be useful for a better understanding of the water flow problem in the soil-plant system in this part of the Northern Great plains, for example, the influence of evapotranspiration to the soil water content; water characteristics of soils from the top 6 in. depth (which has a very dense root distribution) in relation to plant water use; patterns of photosynthetic water use between different species, and even between different modes of physiological activity of one species (Kentucky bluegrass); quantitative data of total leaf area index and its components for selected range sites. These measurements will be repeated in the next couple of years and will be used to fine-tune the plant ecosystem model so that hopefully it can be used to study the dynamic responses for the North Dakota rangeland to natural or human influences.


Acknowledgements

 

We thank Rick Bohn at the CGREC for help in field measurements; Joel Bell and Keith Jacobson at NDSU Department of Soil Science for analyzing soil samples. Finally, we appreciate other CGREC staff members for assistance and support.●

 


Previous section

Table of contents
Data Analysis and Computer Modeling Strategy

  
NDSU Central Grasslands Research Extension Center
Home | Livestock Research | Economics Research | Range Research | 2001 Annual Report