Scientific Investigation

 To build a knowledge base

to support the establishment of sound

public policy

A sufficient knowledge base is required in order for good public policy decisions to be made on water related issues.  The existent knowledge base relating to the Upper Guadalupe River Basin is in need of improvement.  UGRA supports targeted scientific investigation to build the requisite knowledge base.

An example of UGRA's support of scientific investigation is UGRA's providing funding to Texas A & M University the juniper (cedar) interception project and a project examining water use by juniper trees.  Results from the projects can be used in conjunction with simulation models to refine estimate water production from juniper-dominated rangelands in the Edwards Aquifer region.

Another example is UGRA's participation in the Plateau Region Planning Group's Spring Flow Contribution Study.  The goal of the Study is to gain a better understanding of the springs that feed the Upper Guadalupe River.  UGRA has participated in developing the Study's Scope of Work and is compiling existing data for inclusion in the Study.  UGRA will also collect spring-flow data for the Study. 

 Evaporation and Interception Water Loss from Juniper Communities

on the

by

M. Keith Owens

and

Robert K. Lyons

Texas A&M University System

To

Juniper canopies are ideally structured to intercept rainfall and redirect it to the base of the tree, thus altering the hydrology of the site. The amount of water redirected, and the amount lost to interception and evaporation, may be a significant portion of the annual rainfall. We monitored interception and rainfall partitioning in juniper canopies at the Kerr Wildlife Management Area for 1100 days beginning on 3 October 2000. Over this time span, there were 294 total rain events accumulating a total of 70.5 inches of precipitation. Over 170 of the rainfall events (57% of all storms) were less than 0.1 inches and accumulated only 4.9 inches of rainfall. The six largest storms (2% of the total) accumulated over 14 inches of rain. About 35% of the precipitation falling on juniper trees is intercepted by the canopy of the tree and evaporated back to the atmosphere, 5% is intercepted by the litter and duff beneath the tree, and 60% actually reaches the ground surface for either recharge or plant growth. The amount of rainfall intercepted by the canopy is most affected by the intensity and duration of the storm. At high intensities, such as 2.8 inches over a 15 hr period, only 20% is intercepted by the canopy and litter. The remainder is available for either plant growth or aquifer recharge. When rainfall is less intense, such as 0.5 inches over a 19 hour period, 60% (0.3 inch) is intercepted by the tree and litter. Not until nearly 1 inch of rain has fallen does an appreciable amount of water actually reach the ground surface beneath the tree. Nearly 84% of the rainfall events observed over the last several years have been small events of less than 0.5 inches. These events, although common, do not contribute significantly to soil water under juniper trees and are largely ineffective.

Introduction

The density and aerial cover of Ashe juniper (Juniperus ashei) in central Texas has increased over the last 200 years. Originally limited to rocky outcrops or areas of low fuel availability, Ashe juniper now covers almost 2.7 million hectares on the Edwards Plateau.

The impact of juniper trees on the hydrologic budget is hotly debated as water demands from rangelands increase. Understanding both the physiological and physical impact of juniper trees on water availability is crucial; this study investigates the physical impact of juniper trees on the hydrologic budget. The amount of rainfall intercepted by tree canopies and lost to evaporation is species-specific, and may be a function of rainfall intensity (Thurow and Hester 1987, Schowalter 1999, Silva and Rodriguez 2001). When rain falls on a juniper canopy, there are a limited number of things which can happen (Figure 1). The rain can either be intercepted by the juniper canopy or it can fall directly through the canopy to reach the litter layer. The rain that is intercepted can either be evaporated back to the atmosphere or it can flow down the outside of the stem as stemflow. The stemflow water can be further partitioned into water intercepted by the litter layer or water which actually reaches the soil surface. The rain that is not intercepted by the canopy occurs as throughfall and directly reaches the litter layer under the tree. This water is either retained by the litter layer or it can reach the soil surface. It was impractical to follow the rainfall after it reached the soil surface in this study, but it would be available for either plant growth, deep drainage, or overland flow.

Our objectives were to:
1. Determine how rainfall is partitioned within juniper trees at the Kerr Wildlife Management Area, and
2. Determine how rainfall intensity alters the patterns of rainfall partitioning.

The Kerr Wildlife Management Area (KWMA; 30.09 N, 99.49 W) was selected as a  research site on the Edwards Aquifer Drainage Area. Two mature juniper trees were selected for instrumentation in an ungrazed pasture. The trees were representative of the site and were within 100 feet of each other. The site was established on 3 October 2000 and dismantled 1100 days later.

Each tree was instrumented to collect rainfall, throughfall, stemflow and litter moisture (Figure 4). Rainfall above the canopy (hereafter referred to as bulk rainfall) was measured to the closest 0.01 inch using a tipping bucket rain gauge (Texas Electronics).

Throughfall was collected using a system of four 8-inch funnels connected to a collection tube. As the throughfall was collected, a float in the tube recorded the increasing water level. The change in millivolts was calibrated to record the actual height of the water column. After the rain stopped, the datalogger tripped a solenoid to drain the tube and make it ready for the next rainfall event.

Litter moisture was measured using water content reflectometers (Campbell Scientific CS615), after they were calibrated to the high organic matter. The amount of litter was determined by measuring litter depth near the base of the tree, mid-way through the canopy, and at the drip line of the canopy on 8 equally spaced transects radiating from the base of each tree. The area of the tree was combined with litter depths to determine the volume of litter under each tree. Bulk density samples were collected to convert from litter volume to litter mass. Additional samples were taken to calibrate the reflectometer probes. For calibration purposes, the litter was oven-dried and weighed to determine the mass of the sample. Ten percent of that mass was then added using distilled water and a measurement was taken using the CS615 probe. This process was repeated to measure from 10% to 80% gravimetric moisture. This whole process was repeated 6 times and a regression was calculated to convert the millivolt reading from the probes to gravimetric litter moisture. Litter moisture was calculated as :

Litter moisture = -4681.93 + 14416.18* mV - 14600.62 * mV² + 4942.83 * mV³

where mV = millivolt reading from the CS615 probe.

Stemflow was collected by constructing a narrow collar around the base of each tree. The collar collected all of the water which was flowing on the outside of the stem and diverted it to a tipping bucket measuring device. The bucket held 1 L of water before it tipped. The 1 L of water represented about 0.005 inches of rain for an average size juniper tree.

All of this information was collected hourly by an electronic datalogger and downloaded to a computer every second day. The computer then ran a program to check the data for errors and summarized the results, posting the information to a web page at http://uvalde.tamu.edu/intercept.

Canopy interception cannot be measured directly, but must be estimated by subtraction using the formula:

Canopy Interception = Bulk Rainfall - (Throughfall + Stemflow)

And then the amount of water reaching the soil surface was calculated as :

Soil Water = Bulk Rainfall - Canopy Interception - Litter Interception

During the 3-year study, data were collected from over 290 rainfall events. Bulk rainfall was partitioned to canopy interception, evaporation, soil litter interception, and soil moisture, on both a gallons per tree and a percentage basis. Data were analyzed by creating classes of rainfall based on 0.1 inch increments and using curvilinear regression techniques. In addition, the hourly time step of rainfall partitioning for different intensity storms was calculated to determine how rainfall intensity and duration affected interception losses.

Results and Discussion

Tree was 15 feet tall, had a canopy area of 242 square feet, and a litter depth of 1.5 inches. Tree 2 was 14 feet tall, had a canopy area of 220 square feet, and a litter depth of 1.05 inches. These trees are typical of regrowth Ashe juniper after 20-25 years.

Rainfall Distribution
The research site was installed for 1100 days beginning in October 2000. A total of 70.5 inches of rainfall was recorded over this period. It is important to note that this is not the total amount of rainfall received - our equipment could measure only up to 5 inches of rain at a time so larger storms were not used in this report, and equipment malfunction sometimes resulted in missed rainfall. There were 294 individual storms during this interval. When the storms were divided into classes, over 170 of the storms delivered less than 0.1 inch, and most (84%) delivered less than 0.5 inches. Although these storms were numerous, they contributed only 6.8% and 34% of the total rainfall, respectively. The relatively few large events delivered most of the rainfall over this period. Storms greater than 1.25 inches were less numerous (6 total), accounting of only 2% of the total number of storms, but they contributed nearly 20% of the total rainfall. This rainfall distribution will have significant impacts on water availability as demonstrated in a later section.

Rainfall Partitioning
Averaged over all storms during the 3 year study, about 58% of the ambient precipitation reached the soil surface beneath juniper trees while the remaining 42% was intercepted and lost to evaporation. The high canopy interception and evaporative loss is due mainly to the large number of small storms which experienced total, or nearly total, interception. The low intensity storms were numerous but contributed little moisture to the soil surface (Figure 6). Most of the precipitation from storms < 0.1 inch was either intercepted by the canopy (96%) or the litter layer (2%) leaving only 2% of the bulk rainfall to reach the soil surface beneath the juniper trees. At the highest rainfall levels, at least 12% of the bulk rainfall was intercepted by the tree canopy. The litter layer became saturated at fairly low levels of rain and absorbed about 7% of the bulk precipitation, leaving about 81% of the bulk rainfall reaching the soil surface.

As storm size increased, the proportionate amount of water intercepted by the canopy and lost to evaporation decreased (Figure 6). Curvilinear regression analysis demonstrated the high interception loss from small rainfall events. Approximately 50% direct throughfall did not occur until at least 0.4 inches of rain occurred. At this time, about 43% of the rain was intercepted by the canopy, 5.6% was intercepted by the litter and 2% occurred as stemflow. The remaining 50% directly reached the soil surface. At the highest rainfall levels, nearly 88% of the rain directly reached the soil surface as throughfall, nearly 7% was intercepted by the litter layer, 6.7 % occurred as stemflow and 8.7% was intercepted by the canopy. Interception by the litter layer peaked quickly and remained constant after saturation, resulting in a low coefficient of determination for that regression.

Rainfall Partitioning Model
We created a simple model combining average tree size, the frequency distribution of rainfall events, and the regression equations from Figure 6 to calculate the impact of juniper trees on the hydrological budget at each of the 10 research sites. These estimates are based on the solitary trees we measured, although as tree density increases the canopies may influence one another to some extent. We included a range from 20% canopy cover, which would be an open savanna, to 100% canopy cover which represents a cedar break. We made a conservative assumption that all of the bulk rainfall reaches the soil surface in a grassland savanna. When juniper cover was low (20%), the amount of water lost to canopy and litter interception was about 0.2 acre-feet per year, regardless of the site (Figure 7). Intuitively this makes sense because the types of storms and the amount of rainfall should not affect water loss when tree cover is low. As tree cover increased from 20 % to 100%, the amount of water lost to interception increased to an average of 1.05 acre-feet (342,000 gallons) per acre per year. At the KWMA, the pattern of storms resulted in an average canopy interception loss of 0.82 acre feet per acre of cedar break.

Another use of this model is to determine the amount of water which can be gained into the soil when juniper is removed. For instance, if a cedar break at the KWMA site was reduced by 80%, the expected increase in water at the soil surface would be 0.71 acre-feet per year (231,000 gallons). At this point we cannot determine how much of this water would be available for directly recharging the aquifer; that is the objective of another on-going study. The important point is that removing the juniper will result in a net gain of water to the ecosystem. An additional caveat is that vegetation regrowth will also affect the amount of water intercepted by plant canopies as the site recovers.

Rainfall Intensity and Partitioning within Juniper Canopies
Low Intensity Storms.
Low intensity storms typically deposit < 1 inch of rain over a 24 hour period. During low intensity rainfall events, most of the initial rainfall is intercepted by the canopy and the litter layer. Figure 8 depicts the hourly partitioning of rainfall during a 0.5 inch storm that lasted for 29 hours. During the first 16 hours of the storm, canopy interception and litter interception were the dominant factors. After 0.3 inches of rain accumulated (at hour 17), then throughfall became the dominant factor in partitioning rainfall. Overall stemflow was a negligible factor in low intensity storms. The cumulative partitioning (Figure 9) demonstrates that over 50% of the rain received during this a typical low intensity storm is intercepted by either the tree canopy or the litter layer.

High Intensity Storms.
High intensity storms can deposit 1 inch or more over a very short time period. The hourly pattern of rainfall within high intensity events dictates how rainfall is partitioned within tree canopies. Figures 10 and 11 depict a 2.7 inch storm which began with a light rain over a 16 hour period. The hourly time steps (Figure 10) show that periods of low rainfall typically have higher interception losses and lower throughfall. During the first 0.3 inches of the storm, most of the rainfall was captured by either the canopy or the litter (up to hour 3 in Figure 11), but after that throughfall was the dominant factor. Hours within the storms that had high intensity rainfall (for example hours 6 to 8, and 11 to 13) experienced greater throughfall than other periods. Stemflow seemed to lag behind the rainfall by about 1 hour. The cumulative partitioning (Figure 11) demonstrates that only about 30% of the bulk rainfall received during a mixed intensity storm is intercepted by the tree canopy or litter layer. This particular storm started rather gently with only 0.3 inches over a 3 hour period, but more intense storms behaved differently.

Conclusions

The loss of water due to the physiological process of transpiration has been demonstrated in previous studies. This study demonstrates the clear impact of the physical presence of Ashe juniper on water resources. Over a 3 year period, nearly 40% of the ambient rainfall failed to reach the soil surface beneath juniper trees across a broad geographic region. This effectively changed the precipitation range from 24-36 inches to 14-22 inches under juniper trees. A simple model demonstrates that as much as 1 acre-foot of water per year can be intercepted by juniper canopies within a cedar break and then be re-evaporated to the atmosphere.

In small rainfall events, all of the precipitation was intercepted by the juniper canopy. The infrequent, high intensity storms supply most of the water to the ground surface beneath these trees. The hourly pattern of precipitation within a storm altered the partitioning of rainfall. Storms beginning with brief intense rainfall intercept less water than storms beginning with lower intensities. Hourly time steps within a storm closely mimicked the patterns observed for similar-sized isolated storms.

Juniper trees clearly altered the hydrologic budget simply through their physical presence. Low intensity rainfall, which could conceivably benefit the local plant community, was entirely intercepted by the juniper trees. High intensity rainfall supplies the most water to the system and was less influenced by juniper canopies. The re-direction of bulk rainfall to the stem of the tree via stemflow may benefit the tree by concentrating water near the root system, or conversely it may serve to funnel water to preferential flowpaths beneath the trees. An on-going study is investigating the fate of the stemflow water.

Citations

Thurow, T.L. and J.W. Hester. 1997. Hydrological Characteristics In: Taylor, C.A. (Ed.). 1997 Juniper Symposium. Texas Agricultural Experiment Station, The Texas A&M University System. Tech. Rep. 97-1.

Schowalter, T.D. 1999. Throughfall volume and chemistry as affected by precipitation volume, sampling size, and defoliation intensity. Great Basin Nat. 59:79-84.

Silva, I.C. and H.G.Rodriguez 2001. Interception loss, throughfall and stemflow chemistry in pine and oak forests in northeastern Mexico. Tree Physiology 21:1009-113