Fingered Flow in Laboratory Golf Putting Greens

P.A. Nektarios⁽¹⁾, T.S. Steenhuis^(2)(*), A.M. Petrovic⁽³⁾, and J.-Y. Parlange⁽²⁾  

• Title
  • Introduction
  • Materials
  • Results
  • Conclusions
• References
• List of Figures

Abstract

Golf course putting greens are intensively maintained and often contain a sand based root zone and, therefore, need to be managed to prevent nitrate and pesticide contamination and ground and surface water pollution. In this laboratory study, we investigated the water and solute movement through a typical tri-layer United States Golf Association (USGA) putting green profile. Results indicate that even though the soil layers were homogeneous, preferential flow paths were formed, causing a major portion of the soil profile to be bypassed and, thus, increasing the likelihood of groundwater contamination.  

Introduction

Soil compacting is a major limiting factor of turf grass growth on golf courses (Carrow and Petrovic, 1992). In response, to avoid soil compacting and improve turf grass growth, the United States Golf Association (USGA) developed recommendations for putting green profiles that consist of 30 cm of sand with 1 to 5% organic matter at the surface, an optional 5 cm coarse sand as a middle layer, and a 10 cm gravel layer at the bottom including drainage tiles (USGA, 1993). 

Research has shown that water flow through layered coarse textured soil profiles (similar to putting greens) is at a rate much less than the saturated conductivity of finger-like preferential flow patterns caused by an instability in the wetting front (Hill and Parlange, 1972; Raats, 1973; Diment and Watson, 1985; Glass et al., 1988; Hillel and Baker, 1988; Selker et al., 1992). Once a finger is formed, water and solutes will follow the same path during all infiltration events until its 'memory', based on hysteresis of the soil moisture characteristic curve, is lost by saturation or drying (Liu et al., 1995). The sharp interface between wet finger and dry soil can also be explained by hysteresis (Glass et al., 1989b; Liu et al., 1995).

Fingered flow is not restricted to coarse grained soils but also occurs in water repellent soils (Dekker and Ritsema, 1994; Bauters et al., 1998). Water repellent soils are widespread and occur in Florida, Arizona, and California (DeBano, 1969), as well as abroad in New Zealand, England, and Australia (Bond, 1969; McGhie and Posner, 1980) and the Netherlands (Hendrickx et al., 1988; Dekker and Ritsema, 1994). Some researchers even suggest that repellency is the norm rather than the exception (Dekker and Jungerius, 1990; Wallis et al., 1990). Wetting agents are specifically used by golf courses to reduce water repellency in soils.  

Because of the intensive maintenance requirements and high irrigation and chemical use, putting greens are likely candidates for leaching of nitrates and pesticides (Abrams, 1991). Moreover, fingered flow can cause fast transport of water and chemicals (as well as reduce the efficacy of the applied nutrients and pesticides) and invalidates the predictions of many of the current pesticide fate models that are based on the convective-dispersive equation in which the water and solutes statistically sample all pore spaces. The objective of this study was to characterize the extent of preferential fingered flow through a putting green layered soil profile.

Materials and Methods

A golf putting green was recreated within a glass sided slab chamber 50 cm wide, 46 cm high, and 1.2 cm thick. The profile had three layers of silica sand: the top layer consisted of 30 cm of medium sand with particle sizes between 0.3 to 0.43 mm; the middle layer was 5 cm deep with particle sizes between either 0.45 to 0.85 mm or 0.85 and 1.4 mm; and, finally, the bottom layer, with a height of 10 cm, consisted of very coarse sand with grain diameters between 1.4 and 2 mm. Silica sand, which is transparent, was used to aid in the visualization. The chamber was filled with oven dried sand through a hopper and an extension having three randomizing screens. The sand was gravity packed by adding an extra 15 cm of sand above the desired height. The excess sand was removed with a glass tubing attached onto a vacuum pump (Selker et al., 1992).

Water applications were made with a pesticide nozzle tip (TeeJet XR8001⁽⁴⁾) attached to a chain driven shuttle that distributed water and dye uniformly as described by Selker et al. (1992). Two flow rates were used: 2.8 cm h^-1 and 5.8 cm h^-1 (when the middle layer of the coarse sand was substituted with sand having a particle size range of 0.85 to 0.45 mm). The flow rate was manipulated using a delay switch. All experiments were performed in a constant temperature room at 22^oC 1^oC.

The visualization technique of Glass et al. (1989a) was used to observe the evolution of the moisture content field in the entire chamber. The technique consists of supplying one side of the slab chamber with a diffuse light source composed of a bank high output fluorescent light bulbs run by high frequency ballasts (40,000 Hz). Moisture contents are visualized differences in light intensity at the other site. The brighter the location the higher the moisture content for the same sand. The light intensities were recorded with a black and white video camera (Panasonic, TV Camera WV 1850) and a lap recorder (Panasonic, AG 6720) on videotapes with a time resolution of 1/30 sec. The videotape was processed with IRIS Video Image Software⁽⁵⁾ and used to determine the finger velocity, size, and structure determination for artificial coloring of light intensity differences.  

Eight experimental runs with two different irrigation rates were performed with similar results, therefore, only one run of each different irrigation rate is reported. The experiments started with a water application on dry sand. This is similar to a situation just after construction. After the fingered flow field was fully developed and steady state was reached, the irrigation was stopped. After a 24 hour period, water was applied again at the same rate. To aid in the visualization, the water was colored with FD&C Blue Dye #1 at a concentration of 1.0 g/l. After steady state was reached, a switch from dyed to pure water was made. The experiment was continued until all blue dye was removed from the chamber. The light intensities of the dry sand and subsequent infiltration cycles were recorded on tape. At the end of the experiment, the chamber was saturated from the bottom and a final image was recorded. In a separate experiment, the light intensity and moisture content were correlated (Fig. 1) using the method of Bell et al. (1990) and Liu et al. (1993). Moisture content determination with light intensity in the two bottom layers was not possible because of the high amount of light transmitted by the capillary fringe in the top layer that distorted the rest of the picture. The velocity of each finger was calculated from the location of the finger taken from a video frame approximately every 5 min with the IRIS software.    

Results and Discussion

Water Infiltrating into Dry Sand at 2.8 cm/hr:

The distribution of water in the dry golf putting green is shown in Fig. 2. (This situation is similar to the moisture condition just after construction of the putting green.) For easy visualization of the moisture contents in the top layer, light intensities were translated to the colors of red for saturated and dark blue for dry sand. In the top 2 cm of the profile, water moved as a wavy front (Fig. 2a). An induction (or distribution) zone within the top 2 cm was formed, then the wetting front became unstable and a single (primary) finger developed (Fig. 2b). Previous studies have shown when higher flow rates were applied (60 cm h^-1) to sand, several fingers were formed (Selker et al., 1992). Scanning the light intensity in the induction zone (Figs. 2b and 3) revealed that the water content in the induction zone was the lowest at the starting point of the finger. As expected, we found lower moisture contents are related to higher suctions in the induction zone, a potential gradient funnels or directs all the applied water towards the finger.

Figure 2-e	Figure 2-f
Figure 2-g	Figure 2-h
Figure 2-i

Figure 2

Figure 3

When the primary finger reached the first interface between the sand layers, the downward movement was halted, and water started to move sideways forming a capillary fringe (Fig. 2c). At the same time, two new fingers (secondary) began to form in the induction zone (Fig. 2d). Light intensity observations along the axis of the primary finger just before (44 min) and after the primary finger had reached the interface (50 min), showed that the moisture content was the same except from the induction zone where the moisture content was higher (Fig. 4). This increase in the water content and the associated matric potential in the induction zone of the primary finger, allowed the secondary fingers to develop (Fig. 2d).

Figure 4

Because there was more than one path for the water, the flux (water flow rate) through the secondary fingers was smaller than for the primary finger. As a result, the secondary fingers moved slower, were smaller, and had a lower moisture content than the primary fingers (Table 1, Fig. 5). However, for all three fingers the tip was the wettest part of the finger (Fig. 5). The velocity of the wetting front of the primary finger was, on the average, 0.75 cm/min and relatively constant after the first 5 min when the water content of the induction zone did not change anymore. The velocity of secondary fingers became faster as they become longer where the highest velocity was 0.6 cm/min (Fig. 6).

The moisture content of fingers 2 and 3 at the tip was 2.5 and 2 times smaller, respectively, than the moisture content in the tip of the primary finger. Differences were also observed between the two secondary fingers (Table 1) suggesting that the flux through each secondary finger is not the same and depends on the distance from the primary finger.

After both secondary fingers had reached the capillary fringe, two more (tertiary) fingers developed which moved even slower, and were smaller and dryer than the secondary fingers. Precise data could not be taken because of the small amount of light that was transmitted, compared with the capillary fringe.

Water flowed into the second layer as a marginally unstable front (Fig. 2e). Once it reached the interface of the bottom layer water flowed immediately across as a single finger (Fig. 2e). The finger in the third layer was 2 cm wide which was much smaller than the 8 cm wide finger in the top layer. This is consistent with previous findings in which the finger size is inversely proportional to the grain size (Diment and Watson, 1985; Glass et al., 1989a).

Dye Movement in the Wetted Profile (2.8 cm/hr):

In order to examine the solute path after the profile was wetted, blue dye was applied after a 24 hr drainage period when the outflow from the slab chamber was negligible. Only qualitative observations could be made because the rapid-moisture visualization technique was unable to detect differences in moisture in the dyed water. Dye followed the finger pathways created during the first infiltration cycle. In this case, the water moved through the five fingers that were formed earlier. The velocity through each finger was the same, although the amount of water carried by the primary finger might have been more than the remaining fingers because it was brighter. After the dye traveled the first 20 cm in the top layer, the fingers converged towards the middle of the chamber (Fig. 2f), creating a single finger just over the capillary fringe which concentrated and conducted all the flux from the fingers. The point at which the fingers converged and passed into the middle layer was the same location where the first finger broke through into the second layer in the initial infiltration cycle (Fig. 2g). Dye passed immediately through the interface, without spreading laterally.  

The downward velocity of the five fingers was approximately the same and decreased as they reached the capillary fringe above the first interface (Fig. 7). This decrease was caused by the elevated moisture content and the sideways movement of the fingers (Fig. 2f). The velocity increased again after the five fingers merged into one (Fig. 2h), since the flux from five fingers was concentrated into one.

Figure 7

Water Movement Using a Flow Rate of 5.8 cm h^-1:

A similar, but not identical, infiltration pattern developed at the higher flow rate (5.8 cm h^-1) . After the 2 cm induction zone was developed, a single (primary) finger was initiated. When the primary finger reached the interface, a secondary finger formed (Fig. 2i). Because the middle layer was finer sand (0.85-0.48 mm) than for the lower flow rate, the capillary fringe in the fine sand above the first interface was much smaller. When the secondary finger reached the interface, water broke through to the underlying layer at points where each finger contacted the interface (Fig. 2i). New finger formations in the top layer stopped after breakthrough in the second and third layer had occurred. Once the flow pattern was established, the paths became wider but, otherwise, remained the same for the entire duration of the study (one week). Fig. 2j shows the water flow pattern after blue dye was added to the water.

Conclusions

Experiments with a layered profile similar to artificial golf putting greens were conducted in slab chambers to visualize water and solute flow. The results show that thin soil slabs are a useful technique to visualize water and solute movement in golf putting greens. We showed that bypass flow may occur in golf putting greens as a result of instabilities at textural interfaces of homogeneous sands. Once a flow pattern was established, it reoccurred again after water application.  

Since the solutes were bypassing most of the soil matrix, the time for pollutants (pesticides or fertilizers) to reach the outlet of the greens is less than when a convective-dispersive flow is assumed. This reduced time could lead to less biodegradation and higher concentration of turf grass chemicals lower in the profile.  

In these experiments, we did not have any plants or thatch layer present which could modify the pattern in the upper layer. The instabilities at the textural interface will not be affected by the absence or presence of plants or thatch because the instabilities are caused by the properties of the lower layer. Further research is needed to confirm the presence of finger and funnel flow under field conditions, as well as procedures that will limit the formation of fingers to reduce the potential for enhanced groundwater contamination due to solute leaching.

References 

List of Figures 

Figure 1. Correlation between normalized light intensity and volumetric water content for the fine sand. The moisture content and intensity were related as: where is the volumetric moisture content and I_n is the normalized moisture content and can be expressed as:  where I is the light intensity of transmitted light, I_s is the light intensity of the saturated slab chamber, and I_d is the light intensity through the dry chamber.

Figure 2-e	Figure 2-f
Figure 2-g	Figure 2-h
Figure 2-i

Figure 2. Finger development in a slab chamber with an artificial golf putting green for irrigation rates of 2.8 cm/hr and 5.8 cm.hr^-1. a) Single primary finger develops for flow rate of 2.8 cm/hr; b) The induction zone remains stationary while the finger moves downward; c) The primary finger water moves sideways when finger reaches textural interface; d) Two more fingers are formed; e) Flow pattern in next two layers; f) Flow paths are shown with FD&C Blue Dye #1 after irrigation water applied again after 24 hr drainage; g) Water flows in the fingers through the capillary fringe to the previously established flow path in the coarse sand and bottom layers; h) Flow path in coarse and bottom layers; i) Flow patterns established for a flow rate of 5.8 cm h^-1; j) Blue dye in irrigation water applied after a period of 24 hr drainage follows the same pathways created earlier.  

Figure 3. Water content in the induction zone shows a lower moisture content where the primary finger begins. The white dotted line in Fig. 2b indicates the area that was scanned.

Figure 4. Normalized light intensity along the axis of the primary finger at 44 min before the secondary fingers were formed and at 50 min after the secondary fingers were initiated. Notice the differences in the transmitted light intensity in the finger near the induction zone (circled area).

Figure 5. Moisture content along the axis of the dominant finger (Finger 1) and a secondary one (Finger 2) in the fine sand top layer.  

Figure 6. Averaged velocity for the dominant finger (Finger 1) and the two secondary fingers (Fingers 2 and 3).  

Figure 7. Finger location as a function of time for the re-infiltration experiments. At the interface, all water was funneled into Finger 3 and then moved at a fast rate while the other fingers became stationary.

^{1. Consultant, 14 Adersen Str., N. Psychiko, 115 25 Athens, Greece.}

^{2. Professor, Department of Agricultural and Biological Engineering, Riley-Robb Hall, Cornell University, Ithaca, NY 14853; 607-255-2489; Fax: 607-255-4080; E-mail: tss1@cornell.edu.}

^{3. Professor, Department of Floriculture and Ornamental Horticulture, Cornell University, Ithaca, NY 14853.}  

^{*Corresponding Author.}

^{4. Spraying System Co., North Avenue and Schmale Road, Wheaton, IL 60178.}

^{5. 2Data Translation Inc., 100 Locke Drive, Marlboro, MA.}