Soil Science Society of America Journal 66:1344-1349 (2002)
© 2002 Soil Science Society of America
DIVISION S-7—FOREST & RANGE SOILS
Forest Harvesting Influence on Water Table Dynamics in a Florida Flatwoods Landscape
C. M. Bliss and
N. B. Comerford*
Soil and Water Science Dep., Univ. of Florida, 2169 McCarty Hall, P.O. Box 110290, Gainesville, FL 32611-0290
* Corresponding author (nbc{at}mail.ifas.ufl.edu)
 |
ABSTRACT
|
|---|
Flatwoods are an important ecological plant community in the southeastern Coastal Plain with approximately one-third of Florida's flatwoods in timberland. This study was conducted to investigate hydroperiod changes in the water table of both hydric and nonhydric soils following harvest in a Florida cypress [Taxodium distichum (L.)]–pine (Pinus elliottii var. elliottii Engelm.) flatwoods landscape. Shallow water-table wells were established on a 42-ha area and divided into three harvesting treatments containing mature slash pine plantations and pine–cypress swamps: control, clearcut (both hydric and nonhydric soils harvested), and cypress swamps only harvested (hydric soils only harvested). Water-table measurements were obtained at 2-wk intervals for 6 yr. Harvesting treatments occurred 2 yr into the study. Regression equations created from preharvest water-table data between the control block and harvested blocks allowed us to predict uncut water-table responses in the cut areas. These predicted values were compared with the actual postharvest observations. The water-table level increased 48 and 49 cm and 19 and 21 cm in the hydric and nonhydric soils in the clearcut and cypress swamps only cut, respectively, during the first 126 d after harvesting. Significant differences (P < 0.10) in water-table depth because of harvesting occurred throughout the four postharvest years, but the differences were less than seen initially, and the pattern was seasonal. Compared with the predicted uncut condition, water tables tended to be lower (drier) during the growing season and higher (wetter) during the nongrowing season. These seasonal fluctuations were presumably driven by changes in evapotranspiration rates resulting from differences in leaf area of the pine canopy and understory.
|
INTRODUCTION
|
|---|
THE FLATWOODS LANDSCAPE, an intimate mixture of cypress swamps and southern pine stands, is an important ecological plant community of the southeastern Coastal Plain. It occupies 
50% of the Florida land area, or 3 million hectares (Abrahamson and Hartnett, 1990). Approximately 1.9 million hectares are in managed pine plantations, which make up one-third of Florida's timberland (Brown, 1996). The flatwoods are also used for range and other agricultural crops such as citrus and potato.
The surficial water table connects the pine–cypress swamps and pine-plantation components of the flatwoods landscape. Hydric soils, which support pine–cypress swamps, are level to depressional, very poorly to poorly drained, and often inundated areas of these landscapes. These soils are surrounded by nearly level and poorly to somewhat poorly drained, nonhydric soils that often support commercial pine plantations (Keuhl et al., 1997). The proximity of the surficial water table and its hydroperiod, the fluctuation of the water table over time, have a substantial influence on the type and productivity of plant communities (Johnson and Bell, 1976; Glaser et al., 1990; Keddy and Reznicek, 1986).
The water-table hydroperiod is dependent on the balance between rainfall and evapotranspiration, with lateral and subsurface drainage playing a limited role (Ewel and Smith, 1992). Lateral water flow rates have been estimated at 12 cm d-1 (Riekerk, 1992; Crownover et al., 1995) in soils that have hydraulic conductivities as high as 76 cm h-1 when the slope is nearly level. Vertical subsurface drainage is limited by clay layers with hydraulic conductivities of <1 cm h-1 (Tan et al., 1999).
Forest harvesting changes the rate of transpiration and, hence, water-table dynamics (Riekerk, 1989; Sun et al., 2000, 2001). Immediately after harvesting, leaf area is drastically reduced, causing the water table to rise (Trousdell and Hoover, 1955; Williams and Lipscamb, 1981). The water table has been shown to increase significantly after clearcutting or selective harvesting in coastal plain systems (Trousdell and Hoover, 1955; Williams and Lipscamb, 1981; Sun et al., 2000, 2001). However, intensive management such as fertilization, weed control, and improved site preparation for the regeneration and growth of the subsequent stand could potentially increase leaf area and water use over the original levels, resulting in a changed water-table hydroperiod (Xu et al., 1999, 2000). An important concern is whether this will cause water tables to eventually become lower than preharvest levels.
Flatwoods ecosystems are influenced by the shallow fluctuating water table, yet there is little detailed information on the water-table hydroperiod and the influence that forest harvesting and site regeneration have on the hydroperiod. The objective of this study was to document the change in the hydroperiod of the surficial water table of a typical flatwoods managed ecosystem during the 4 yr following forest harvesting.
|
MATERIALS AND METHODS
|
|---|
Study Area
The study area was located 33 km northeast of Gainesville, Florida. Approximately 31% of the 42-ha area was in hydric soil swamps (Kuehl et al., 1997) while the remaining area was a managed slash pine plantation underlain by sandy, siliceous, thermic Ultic Alaquods (Pomona fine sand and sand). From April 1992 through May 1998, the average annual temperature was 20°C, with a mean monthly low of 6°C in January and a mean monthly high of 33°C in July. Average yearly rainfall was 1150 mm yr-1 (data taken from Green Acres Agronomy Research Facility, located 32 km northwest of Gainesville). There are two distinct dry periods within a year with the first dry period being from approximately April to June and the second from October to December. Rainfall data for this study were recorded daily at a weather station located at the study site (Fig. 1)
. Total monthly, seasonal, and yearly rainfall is in Table 1. The growing season for the pine plantations was defined as March through October (Neary et al., 1985).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1. Rainfall for the entire study period. Each bar represents accumulation from the previous measurement date (2 wk intervals).
|
|
View this table:
[in this window]
[in a new window]
|
Table 1. Annual and seasonal rainfall amounts measured at Green Acres Agronomy Research Facility, near Gainesville, Florida.
|
|
Establishment of Water-table Wells and Well Measurement
In January 1990, the study area was surveyed to establish a 50 by 50 m grid system. Water-table wells were established at approximately every second grid point for a total of 120 wells. Five-centimeter diameter, 1.5-m long polyvinyl Cl water-table wells were installed using a 5-cm hand auger. Well depths varied from 1 to 1.4 m, depending on the depth of the argillic horizon and water-table conditions at the time of installation. The bottom 1 m of the pipe had well screening attached to a well point (Crownover et al., 1995). Well locations are illustrated in Fig. 2
. A more detailed description of the establishment of the wells and reports on earlier data can be found in Crownover et al. (1995), Comerford et al. (1996), and Keuhl et al. (1997).
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 2. Map of the study site with water-table wells and cypress swamps marked. The location of cypress swamps roughly corresponds with location of hydric soils; outside the swamps tend to be nonhydric soils.
|
|
Water-table depth was measured on a biweekly schedule beginning 6 Apr. 1992 and ending 31 Mar. 1998 because of a wildfire that destroyed the study site. A flashlight attached to a retractable tape was used to measure water depth to the nearest centimeter. The depth of the water table below the soil surface was determined by combining the water depth in the well and the well height above the surveyed grid point. A more detailed description of these methods is given in Crownover et al. (1995).
Harvesting Treatments and Water-table Measurement
In April and May of 1994, two silvicultural treatments were administered to the research area. The southeast (SE) block of the study area was clear-cut (clear-cut), including both the pine plantation and the pine–cypress swamps. In the northwest (NW) section, only the cypress swamps were harvested (cypress-cut). The southwest (SW) section was left as a control (control; Fig. 2). In September and November 1994, the SE block was bedded. In January 1995, that area was planted with slash pine, and in February 1995, 437 ml of Arsenal (C13H15N3O3·C3H9N) per hectare was applied in 1.5-m wide bands down the planting row to control broad leaf vegetation. The NW block was allowed to regenerate naturally.
The soil at each water-table well was classified as hydric or nonhydric in a previous study conducted by Keuhl et al. (1997). Classification was based on soil wetness indicators of the water-table depth (Keuhl et al., 1997). By using the water-table depth data prior to harvest (April 1992 to April 1994), regression equations for each well in the clear-cut and cypress-cut blocks were developed relative to wells in the control block. The purpose of these regressions was to predict water-table depth for the unharvested condition of those wells in the harvested blocks. Wells in the control block that had correlation coefficients >0.90 (generally >0.95) with each well in the clear-cut and cypress-cut blocks were identified, and the regression coefficients were calculated. Coefficients for up to three wells in the control block were identified for each well in the harvested blocks. For each measurement date, wells in the control block were used to predict what the water-table depth would have been assuming harvesting had not occurred. Examples of the predicted water-table levels versus the observed levels using the regression equations are shown in Table 2. The error percentage for one of the driest dates was 4, 9, 2, and 2 for the cypress-cut hydric soils, cypress-cut nonhydric soils, clear-cut hydric soils, and clear-cut nonhydric soils, respectively. On one of the wettest days the percentage of error was 4, 5, 7, and 3 for the cypress-cut hydric soils, cypress-cut nonhydric soils, clear-cut hydric soils, and clear-cut nonhydric soils, respectively.
View this table:
[in this window]
[in a new window]
|
Table 2. Examples of results of regression analysis (slopes and standard errors) of the measured versus predicted water-table depths using preharvest data from the harvested blocks (October 1993–March 1994).
|
|
Statistical Analysis
The predicted water-table depth for each well was compared with the actual water-table measurements in the harvested SE and NW blocks. The difference (observed - predicted) was assumed to be the water-table response to harvesting. Wells along the borders of the treatment blocks were not used because of a possible edge effect on the water table.
Differences between values of the predicted uncut condition and observed values of the wells in the harvesting treatments were calculated for all sampling dates. During dry periods, if >25% of the wells were dry (water-table depth below the depth of the measurement well), data for that day were excluded from the analysis. A value of zero for the difference between the observed water-table depth and the predicted water-table depth indicated that the water-table depth is at the preharvest level. A positive value indicated that the observed water table was closer to the soil surface than it would have been if it had not been harvested. A negative value indicated that the observed water table was lower that it would have been if not harvested. Statistical analysis was conducted on the differences (predicted - observed) to determine if water-table levels at each date were significantly different (P < 0.10) from zero, indicating a water-table response to harvesting. The repeated measures analysis of variance model was used to determine significantly different dates. This model is similar to a three-way ANOVA model by using dates, hydric or nonhydric, well location, and their interactions, except that correlated error terms were allowed.
|
RESULTS
|
|---|
Harvesting occurred in the spring of 1994, when the water table was dropping (Fig. 3)
. In the control block, the average water-table depth dropped from -0.10 m to more than -0.60 m just prior to harvesting. For 4 mo after harvesting, Day 891 through Day 1017, water tables under both harvesting regimes increased in all landscape positions (Fig. 4 and 5)
. The clear-cut block, showed the greatest response to harvesting. The water table was up to 48 cm higher in the hydric soils and 49 cm higher in the nonhydric soils (Fig. 4). The water table in the cypress-cut block had less of a response, increasing just 21 cm in the hydric soils and 19 cm in the nonhydric soils (Fig. 5).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 3. Average water table depth at each 2-wk collection date for the control block (both hydric and nonhydric soils) for the entire study period. The arrow indicates when harvest occurred. Breaks in the hydroperiod line represent times when the water table was below the depth of the water table wells.
|
|
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 4. Differences between observed minus predicted water-table levels over time following harvesting (June 1994–March 1998) for the clear-cut block. The upper line represents the hydric soils and the lower line represents the nonhydric soils. Grey shaded areas indicate the growing season. Dates significantly different from zero are indicated by an ‘*’. Missing data indicate dry wells.
|
|
View larger version (51K):
[in this window]
[in a new window]
|
Fig. 5. Differences between observed minus predicted water-table levels over time following harvesting (June 1994–March 1998) for the cypress-cut block. The upper line represents the hydric soils and the lower line represents the nonhydric soils. Grey shaded areas indicate the growing season. Dates significantly different from zero are indicated by an ‘*’. Missing data indicate dry wells.
|
|
After the first 126 d (beyond Day 1017, October 1994) the influence of harvesting on the water table diminished and stabilized. While there were still significant effects on the water-table depth over time, they rarely were of the magnitude seen during the first 4 mo. The water-table response in the clear-cut block ranged between 25 cm lower to 25 cm higher than the predicted uncut condition. In the cypress-cut block, the water-table response ranged from 10 cm higher to 10 cm lower than the predicted uncut condition (Fig. 4 and 5).
The hydric soils, which were harvested in both treatments, exhibited similar water-table hydroperiods (Fig. 4 and 5). Fifty-one percent (clear-cut block) and 67% (cypress-cut block) of the measurement dates during the growing season were not significantly influenced by harvesting. Of the dates that were influenced by harvesting, 20 out of 23 (87%) and 12 out of 17 (71%) had water tables lower than the predicted uncut condition for the clear-cut and cypress-cut blocks, respectively (Fig. 4 and 5, Table 3). During the nongrowing season, 72% (clear-cut) and 83% (cypress-cut) of the total measurement dates were not influenced by harvesting. When a significant difference in the water-table level was encountered, 7 of 8 d (88%) in the clear-cut block and 3 of 5 d (60%) in the cypress-cut block were higher than the predicted uncut condition (Fig. 4 and 5, Table 3).
View this table:
[in this window]
[in a new window]
|
Table 3. Number of measurement dates from Day 1017 to the end of the study with water-table levels that are not significantly different, have a significantly higher water table, and have a significantly lower water table compared with the uncut condition. Significantly different dates indicate area is still being affected by harvesting. The number of dry wells is also included. The Cypress-cut block had only cypress swamps (hydric soils) cut and in the Clear-cut block both swamps and pine plantations (hydric and nonhydric soils) were clearcut.
|
|
When comparing the nonhydric soils from both harvesting treatments, we found that the water-table levels in the two harvesting treatments reacted differently. The cypress-cut block, with an uncut slash pine plantation still occupying the site, had 79% of all measurement dates (75% of the growing season and 90% of the nongrowing season) unaffected by the treatment, while 11% of the time it was too dry to get measurements. In contrast, the water table in the clear-cut block was not affected by harvesting only 24% of the time (29% of the growing season and 14% of the nongrowing season), while 12% of the time the water table was too low to measure (Fig. 4 and 5, Table 3).
The response of the nonhydric soils in the clear-cut block changed with time after harvest. During the growing season, seven of nine (78%)of measurement dates in 1995, five of ten (50%) in 1996, and two of six (33%) in 1997 had water tables higher than the uncut condition (Fig. 4). During the nongrowing season, all measurement dates in the clear-cut block that were significantly different than the predicted uncut condition (83% of the measurement dates) were characterized by water tables higher than the uncut condition.
|
DISCUSSION
|
|---|
The immediate decrease in water-table depth after harvesting has been documented by Trousdell and Hoover (1955), Langdon and Trousdell (1978), and Williams and Lipscamb (1981) in other southern locations. This result was also previously reported in a study from this site (Sun et al., 2000) based on three wells in hydric soils and three wells in nonhydric soils that are in close proximity to hydric soils. The current data, based on 115 wells, show that the results of Sun et al. (2000) extend over the entire landscape. The increase in water yield and movement to surface waters because of both harvesting treatments was significant for only 4 mo. The degree and magnitude of the increase were related to harvest treatment, time of the year, and the amount of precipitation received during this period.
Four months after harvest, water-table levels returned to a normal hydroperiod. These data show that any variation from that normal hydroperiod was related to harvesting treatment and landscape position (hydric vs. nonhydric). Our data shows that in 21 to 36% of the growing season measurements during the 4 yr following harvesting, hydric soils had water tables 5 to 9 cm lower than if harvesting had not occurred. This is assumed to be because of vegetation regrowth and regeneration. Ewel (1985) showed that regeneration of vegetation in cutover cypress swamps had increased levels of evapotranspiration compared with uncut areas. A study by Aust et al. (1997) in Alabama found a similar result with a lower water table during the growing season 7 yr after harvesting because of regrowth of vegetation. The length of significant harvesting affect and the decreases in water yield across the landscape is a key component of predicting cypress–pine flatwoods hydroperiods. Sun et al. (2000) also found harvesting significantly affected the water-table levels during the growing seasons 2 yr after harvesting, but not during the nongrowing season as we did. The disagreement is probably because of the fact that our study used 115 wells distributed over a 42-ha landscape, while Sun et al. (2000) based their conclusions on six wells, three in cypress swamps and three outside, but in proximity to cypress swamps.
The reaction of the water table in nonhydric soils differed from hydric soils. Nonhydric soils of the cypress-cut block were not harvested. Ninety percent of the time measurements showed they were not different from the predicted uncut condition, thus giving one confidence in the regression equation's ability to predict the uncut condition. In contrast, the forest stand on the nonhydric soils of the clear-cut block was harvested and 64% of the time during the 4 yr following harvesting the water-table level differed from the predicted uncut condition. As the new forest stand and understory vegetation regenerated, the tendency was for the water table to go from higher (wetter) than the predicted uncut condition because of a decrease in leaf area and lower evapotranspiration following cutting, to drier (lower).
While leaf area and evapotranspiration were not measured during the 4-yr period, the original stand had a low stocking (5 m2 basal area), and so had a low leaf area index. Site preparation, improved genetic planting stock, and early weed control are known to accelerate early stand development on these soils (Neary et al., 1985). As the new stand developed, the water-table level during the growing season changed from being 4 to 12 cm higher than the predicted uncut condition to being 10 cm drier for limited periods of time, suggesting some degree of increased water use compared with the previous stand. Interpretation of these changes is complicated by the tendency of the water table to be higher during the nongrowing season. This study was not designed to model the effect of these seasonal changes at the landscape level and determine what impact it would have on water yield, but a landscape level approach to examine this trend is the key to determining the effect of forest harvesting and improved forest management on water yield.
These results show that forest management can influence water-table hydroperiod for extended periods of time after harvesting treatments. Water yield from cutover forest stands on coastal plain Spodosols will increase immediately after harvesting. However, within a matter of months the influence of harvesting on the water-table depth is diminished. For 4 yr after harvest, appreciable parts of both the growing and nongrowing seasons have periods where the water table is still influenced by the changing vegetation dynamics initiated by the harvest. The net effect on water yield from a landscape depends on the area of the landscape affected by harvesting, vegetation present, and the magnitude of water-table changes following harvesting. This evaluation would be possible with a landscape level–process model approach that incorporates these data. Our data indicate that such an evaluation is necessary to understand the management implication of harvesting on water dynamics in forested Spodosols of the lower Coastal Plain.
|
ACKNOWLEDGMENTS
|
|---|
The authors wish to gratefully recognize the work and financial support of the National Council of Air and Stream Improvement, in particular Dr. James Shepard. In addition, the assistance of Dr. Dan Neary and the U.S. Forest Service is recognized for their help in the initial years of the study. The study would not be possible without the help of The Timber Co. on whose land the study was established and who did all forest management at the area during the study. This paper is Florida Agricultural Experiment Station Journal Series Number R-08507.
Received for publication July 20, 2001.
|
REFERENCES
|
|---|
- Abrahamson, W.G., and D.C. Hartnett. 1990. Pine flatwoods and dry prairies. p. 103–149. In R.L Myers and J.J. Ewel (ed.) Ecosystems of Florida. Univ. Central Florida Press, Orlando, FL.
- Aust, W.M., S.H. Schoenholtz, T.W. Zaebst, and B.A. Szabo. 1997. Recovery status of a tupelo-cypress wetland seven years after disturbance: Silviculteral implications. For. Ecol. Manage. 90:161–169.
- Brown, M.J. 1996. Forest statistics for Florida, 1995. United States Department of Agriculture-Forest Service. Southern Research Station. Resource Bulletin SRS-6, USDA-Forest Service, Asheville, NC.
- Comerford, N.B., A. Jerez, A.A. Freitas, and J. Montgomery. 1996. Soil water table, reducing conditions, and hydrologic regime in a Florida flatwood landscape. Soil Sci. 161:194–199.
- Crownover, S.H., N.B. Comerford, D.G. Neary, and J. Montgomery. 1995. Waterflow patterns through a cypress/pine flatwoods landscape. Soil Sci. Soc. Am. J. 59:1199–1206.[Abstract/Free Full Text]
- Ewel, K.C. 1985. Effects of harvesting cypress swamps on water quality and quantity. Florida Water Resources Research Center. Publication no. 87. Florida Water Resources Research Center, Gainesville, FL.
- Ewel, K.C., and J.E. Smith. 1992. Evapotranspiration from Florida pond cypress swamps. Water Resour. Bull. 28:299–304.
- Glaser, P.H., J.A. Janssens, and D.I. Siegel 1990.The response of vegetation to chemical and hydrological gradients in the Lost River peatland, northern Minnesota. J. Ecol. 78:1021–1048.
- Johnson, F.L., and D.T. Bell. 1976. Plant biomass and net primary production along a flood frequency gradient in a streamside forest. Castanea 41:156–165
- Keddy, P.A., and A.A. Reznicek. 1986. Great Lakes vegetation dynamics: The role of fluctuating water levels and buried seeds. J. Great Lakes Res. 12:25–36.
- Kuehl, R.J., N.B. Comerford, and R.B. Brown. 1997. Aquods and Psammaquents: Problems in hydric soil identification. p. 41–59. In M.J. Vepraskas and S.W. Sprecher (ed.) Aquic conditions and hydric soils: The problem soils. SSSA Spec. Publ. 50. SSSA, Madison, WI.
- Langdon, O.G., and K.B. Tousdell. 1978. Stand manipulation: Effects on soil moisture and tree growth in southern pine and pine hardwood stands. p. 221–236. In W.A. Balmer (ed.) Soil Moisture and Site Productivity Symposium. USDA Forest Service, Southeastern Area, Atlanta, GA.
- Neary, D.G., T.E. Cooksey, N.B. Comerford, and P.B. Bush. 1985. Slash and loblolly pine response to weed control and fertilization at establishment. Proc. South. Weed Sci. Soc. 3:246–253.
- Riekerk, H. 1989. Infleunce of silvicultural practices on the hydrology of pine flatwoods in Florida. Water Resour. Res. 25:713–719.
- Riekerk, H. 1992. Groundwater movement between pine uplands and cypress wetlands. p. 644–654. In M.C. Landin (ed.) Wetlands, Proceedings of the 13th Annual Conference, Society of Wetland Scientists, New Orleans, LA.
- Sun, G., H. Riekerk, and L.V. Kornhak. 2000. Ground-water–table rise after forest harvesting on cypress-pine flatwoods in Florida. Wetlands 20:101–112.
- Sun, G., S.G. McNulty, J.P. Shepard, D.M Amatya, H. Riekerk, N.B. Comerford, R.W. Skaggs, and L. Swift, Jr. 2001. Effects of timber management on wetland hydrology in the eastern United States. For. Ecol. Manage. 143:227–236.
- Tan, Z., W.G. Harris, and R.S. Mansell. 1999. Water table dynamics across an Aquod-Udult transition in Florida flatwoods. Soil Sci. 164:10–17.
- Trousdell, K.B., and M.D. Hoover. 1955. A change in ground-water level after clearcutting of loblolly pine in the Coastal Plain. J. For. 53:493–498.
- Williams, T.M., and D.J. Lipscomb. 1981. Water table rise after cutting on Coastal Plain soils. S. J. Appl. For. 5:46–48.
- Xu, Y.J., W.M Aust, J.A. Burger, S.C. Patterson, and M. Miwa. 1999. Recovery of hydroperiod after timber harvesting in a forested wetland. USDA For. Ser. Gen. Tech. SRS-30:282–287.
- Xu, Y.J., J.A. Burger, W.M Aust, and S.C. Patterson. 2000. Responses of surface hydrology and early loblolly pine growth to soil disturbance and site preparation in a lower coastal plain wetland. N.Z. J. For. Sci. 30:250–265.
TOP
ABSTRACT
INTRODUCTION
