0237-B1

Peak Flow Recovery after Clear Cutting, Quebec and Sâo Paulo

François Guillemette, André P. Plamondon^[1], Walter P. Lima, Maria José B. Zakia and Denis Lévesque

Abstract

An approach is proposed to develop a generalized peak flow recovery curve after forest harvesting. Rainfall peak flow recovery rates, after clear-cutting a basin in the boreal forest at Montmorency forest (Québec, Canada), were compared with values obtained after clearing and planting eucalyptus on two basins at Bela Vista (Sâo Paulo, Brazil) where tree growth is much faster. Both basins in Bela Vista had similar peak flow recovery rates and had fully recovered after six years at tree height of 12 m. At Montmorency forest, the seven-year period of peak flow recovery, while tree height reached only 1 m, was not sufficient to establish a recovery curve against tree height. However, most of the rainfall and snowmelt annual recovery values fluctuated below the curves for Bela Vista, indicating that the hypothesis of a generalized relation is worth pursuing. This is further supported by the similarity of Bela Vista curves with the snowmelt recovery rates at Montmorency forest, which are considered a good index of peak flow recovery. It is suggested to validate this approach using recovery rates for longer periods and on other basins.

Introduction

Basin studies, reviewed by Plamondon (1993) and MacDonald et al. (1997) indicate that peak flows can be increased by tree harvesting. Our recent interest for peak flow comes from the presumption that an increase, particularly at bank-full discharge can modify streambed morphology (Dunne and Leopold 1978) and, consequently, the aquatic habitat.

The effect of harvesting on peak flow decreases with regeneration establishment (Hornbeck et al. 1997; Thomas and Megahan 1998; Verry et al. 1983). This effect is expressed by a peak flow augmentation coefficient (PFAC = 100 % - hydrologic recovery) in this paper. The PFAC enables the forester to calculate the equivalent clearcut area (ECA) of previous cuts and to set the maximum area for the next intervention to satisfy a pre-established ECA (ex. 50 %). The best way to obtain the PFAC is from basin studies, but this long term (15-20 years in boreal forest) approach is expensive. It was then supposed that the hydrologic recovery of peak flow is linked to vegetation growth, as was also observed for annual water yield (Swank et al. 1988). It was hypothesized that a generalized relation could be established between the PFAC and the vegetation growth, permitting the use of the same relation in various forest ecosystems. Thus, the purpose of this study was to compare the PFAC for a basin (7A) in the humid microthermal climate (Köppen) of Montmorency Forest to the ones of two basins (A and B) in the humid (dry winter) mesothermal climate of Bela Vista where tree growth is much faster.

Sites and methods

Montmorency forest

The Montmorency Forest (47°N, 71°W) is located 80 km north of Quebec City. The average annual temperature, precipitation and snowfall water equivalent are respectively 0.3 °C, 1416 mm and 465 mm. The top 30 cm of soil is highly permeable while the underlying till has a low permeability. This coniferous boreal forest is dominated (75 % of the volume) by balsam fir (Abies balsamea (L.) Mill). One year after logging, herbaceous plants and shrubs (Rubus idaeus L., Ribes triste Pallas. and Prunus pensylvanica L.f.) along with the commercial species covered 80-100 % of the ground. The commercial volume (dbh > 9 cm) of balsam fir stands normally vary between 75 and 197 m³ha^-1 at age 60.

The effect of clearcutting basin 7A on instantaneous (hourly) peak flow was evaluated using the paired basins approach with a calibration period (Table 1). Streamflow was obtained with V-notch weirs. Since patch cutting basin 6 did not have a significant effect on peak flows (Plamondon et al. 1998), it was used as control (1985-92) for the treatment on basin 7A. Problems in using a former treated basin as a control have been discussed by Thomas (1990).

Table 1 Treatments on basins 6 and 7A.

The changes of rainfall peak flow (Qpk), were evaluated for each storm corresponding to a daily streamflow above 0.5 m³ s^-1 from a 917 ha watershed in which basins 6 and 7A are nested. All events without snow on the ground and with comparable total rainfall (within 25 %) were used. For multiple-peak events, the highest peak on the basin 6 was used. Events were separated by a fall in the discharge reaching a constant slope separation line as used by Plamondon et al. (1998). Since the highest peak flow increased occurred while 60 % to 85 % of the area of basin 7A was harvested, ten events of summer 1994 were included in the analysis. Streamflow data of summer 1996 were rejected due to sand accumulation at the gauging station. Peak flows were log transformed to meet the assumption of homoscedasticity and to improve the frequency distribution of the data along the interval of the regression. Linear models developed for the period of calibration and the post-treatment periods were compared with the use of a dummy variable for the effect of clearcutting on the intersect. Multivariate linear regression (GLM, SAS Inst.) with a dummy variable to evaluate the treatment effect on the intercept alone can be applied since we know (Guillemette et al. 2002) that clearcutting at the Montmorency Forest does not affect significantly the slope of the regression for peak flows with a return period up to 5 years. Moreover, this method is more appropriate when the sample size is small, since it reduces the possibility for changing the regression slope. The peak flow augmentation coefficients (PFAC) were calculated (i) firstly as the ratio of the mean annual peak flow increase on the mean of 1994 and, (ii) secondly as the ratio of the increases of the maximum annual peak flow to the 1994 maximum. The first year’s mean and maximum increases of 47 % and 63% respectively corresponded to 100 % PFAC.

Bela Vista

The Bela Vista experimental basins are located (23º25’S, 45º54’W), 79 km south-east of Sao Paulo, at an altitude of 700 m (Vital et al. 1999a). Average maximum and minimum temperatures are respectively 23 and 17 ºC. For the 1987-95 period, annual precipitation was 1328 mm. Annual water yield has decreased from about 250 and 400 mm, for basins A (7 ha) and B (6.5 ha) respectively, during 1987, to less than 100 mm on each basin in 1994. The regional geomorphology is characterized by scarps with a topography of prolonged valleys with rectilinear paths of drainage. The colluvial/alluvial deposit on basin A has allowed for the development of a softer topography, deeper soil, smaller textural gradient between horizons A and B, greater percentage of flocculated clay and higher infiltration capacity when compared to basin B. ^[2]In April 1987, both basins were planted (2 m × 3 m) with Eucalyptus saligna (Smith.) following harvesting of the dense ombrophilous forest, prescribed burning and soil preparation. Basin A (slope 19.6 %) was prepared for plantation with disc scarification to a soil depth of 20 cm. Plantation holes of 30 cm × 30 cm × 30 cm were dug on basin B (slope 28.9 %). The commercial volume of eucalyptus reaches 178 m³ha^-1 at harvesting time, 7 years after planting.

The treatments effects on instantaneous peak flow, for rainfall above 30 mm, were evaluated on both reforested basins (A and B) without a calibration period or a control basin. The sixth year after planting (1992-93) the maximum peak flow was more than 900 times smaller than during year 1 and was used as the full hydrologic recovery. Mean annual and the 20 % highest peak flows of the first five years were compared to those of the sixth year. This variable is less influenced by a single extreme event. The first year’s annual increase (PFAC of 100 %) for the 20 % highest peak flow on basins A and B were 733 % and 2800 % respectively. Since the PFAC calculated from the mean annual peak flow were very similar to those calculated from the 20 % highest peak flow, only this last variable was retained.

Results and discussion

Clearcutting 85 % of basin 7A had a significant (p<0.100) effect on peak flow for any year during the post-treatment period (Table 2). Data for the years 1997 and 1998 were merged to increase the sample size and to improve the distribution of data along the regression line. It is suggested that the peak flow increases are more related to flow path changes due to skid trails and roads than to changes of soil moisture and evapotranspiration. The annual water yield increase of 46 mm (5 % of flow) after cutting was quite small to explain a change of peak flow due to soil moisture. Former measurements in cutovers of basin 6 indicated that the mineral soil water content was not significantly higher than in the mature forest. These small changes were attributed to the wet summers and the presence of herbaceous plants and shrubs quickly invading recent clearcuts. One year after logging basin 7A, herbaceous plants and shrubs along with the commercial species and woody debris, covered 80-100 % of the ground. The absence of a clear peak flow decrease during the first 6 years after harvesting is attributed to the small change of vegetation cover during that period and the slow plant establishment in skid trails and road ditches.

Table 2 Mean annual effect of clearcutting basin 7A (Montmorency Forest) on peak flow.

* Tree height 1997 = 0.41 m; 1998 = 0.68 m.

Eucalyptus growth after planting basins A and B had significantly decreased peak flow (Table 3, Figs. 1A and 1B). The higher than expected peak flow (Fig. 1A) on basin A in 1991 (55^th month) cannot be explained; the hydrograph was clearly visible on the chart. The peak decrease is supported by the one of annual flow observed by Vital et al. (1999b). Clearing on basins A and B caused an important reduction in evapotranspiration, which resulted in higher soil water content before the rainfall events enhancing peak flows. At this site, it is suggested that change of evapotranspiration played a more important role than the soil disturbances on peak flow changes.

Table 3 Annual peak flow increases (%), augmentation coefficients (%) and tree height on basins A and B (Bela Vista).

Fig. 1 Peak flows during post-treatment period on basins A and B (Bela Vista), June 1987 - May 1993.

As expected the hydrologic recovery at Bela Vista was faster than the one in cooler climates. The 8 % annual decrease of the 27 % peak flow increase, following conifers harvesting at Caspar Creek California (Lewis et al. 2001), would yield complete recovery in 13 years. Twenty years after clearcutting basin Basin 1 at H.J. Andrews forest in Oregon peak flows had not completely recovered (Thomas and Megahan 1998) and the effect should last 30 years. In Hubbard Brook, New Hampshire, whole-tree harvesting of northern hardwoods increased peak flows in the first five growing seasons (Hornbeck et al. 1997). Between the 6^th and 12^th years after harvest, only 2 out of 24 peak flows were significantly increased. The relatively small impact of harvesting on peak flows and on the subsequent quick hydrologic recovery were attributed to rapid growth of the regenerating hardwood species.

To compare the hydrologic recovery of biomes with different growth rates the augmentation coefficients (PFAC) for basins 7A, A and B were calculated and plotted against tree height (Fig. 2). Both basins in Bela Vista have a similar peak flow recovery from clearcutting although the peak flow increases were different. From the observations on Bela Vista basins, a PFAC of 30 % was obtained at tree height of 4 m. Peak flows on basin 7A did not recovered over the period 1994-2000, but tree height reached only 1 m. Moreover, because of the important variations of PFAC on basin 7A, few more years would be necessary to draw a general conclusion on the recovery rate of peak flows for these two highly different areas.

Fig. 2 Snowmelt and peak flow augmentation coefficients in relation to tree height.

A study of snowmelt rate under different forest covers at Montmorency Forest (Talbot and Plamondon 2002) yielded snowmelt augmentation coefficient (SAC equivalent to PFAC for snowmelt) of 0 %, or full recovery, at stand height of 10 m corresponding to 67 % of the height at maturity. Furthermore, a mean SAC value of 50 % was obtained at tree height of 4 m. These SAC are similar to PFAC expressed as the tree height at Bela Vista (Fig. 2). The PFAC values calculated from the rainfall peak flow from basin Marcell in Minnesota (Verry et al. 1983) are also very closed to the ones obtained at Bela Vista. However, the snowmelt peak flows from Marcell basin do not appear to have decreased 8 years after clearcutting.

Conclusion

Six years after clearcutting on basin 7A, conifer seedlings have reached about 7% of the height of mature balsam fir stands and rainfall peak flows have not recovered yet. This is consistent with the recovery curve of peak flows on Bela Vista basins, although few more years would be necessary to draw a general conclusion on the recovery rate of peak flows for these two highly different areas. The approach proposed will have to be tested over a longer time and on more watersheds to demonstrate its applicability and limits. Indeed, the post-treatment period on basin 7A represents only a small proportion of the expected recovery period, when compared to basins in Brazil. Clearcutting effects on peak flows are known to be variable between watersheds, while the variability of recovery rates has not been widely studied yet. Thus, arises the need to test this model on more watersheds and forest biomes.

Acknowledgements

This work was supported by the Research Council of Canada (NSERC), the Fonds forestier of Quebec and the Departemento de Ciências Florestais, Universidade de Sâo Paulo.

References

Dunne, T. and L.B. Leopold, 1978. Water in Environmental Planning. Freeman & Company, San Francisco.

Guillemette, F., A.P. Plamondon, D. Lévesque and M. Prévost, 2002. Evaluation of rainfall peak flow response to clearcutting. Submitted to J. Hydrol.

Hornbeck, J.W., C.W. Martin, and C. Eagar, 1997. Summary of water yield experiments at Hubbard Brook Experiment Forest, New Hampshire. Can. J. For. Res. 27: 2043-2052.

Lewis, J., S.R. Mori, E.T. Keppeler and R.R. Ziemer, 2001. Impacts of logging on storm flows, flow volumes and suspended sediment loads in Caspar Creek, California. Am. Geophys. Union, Water Sci. and Appl. Vol. 2:85-125.

MacDonald, L.H., E.E. Wohl and S.W. Madsen, 1997. Validation of water yield thresholds on the Kootenai National Forest. Dep. Earth Resour., Colorado State Univ., Fort Collins.

Plamondon, A.P., 1993. Influence des coupes forestières sur le régime d'écoulement de l'eau et sa qualité: revue de la littérature. Rapport C-47 Ministère des Forêts du Québec.

Plamondon, A.P., D. Lévesque, Y. Ma and M. Prévost, 1998. Long-term effects of forest mosaic management on storm and snowmelt flow, Quebec. Proc. British Hydrol. Society John Wiley & Sons, Vol. 1:503-515.

Swank, W.T., L.W. Swift and J.E. Douglass, 1988. Streamflow changes associated with forest cutting, species conversions, and natural disturbances. Ecol. Studies 66: 297-312.

Talbot, J. and A.P. Plamondon, 2002. The diminution of snowmelt rate with forest regrowth as an index of peak flow hydrological recovery, Montmorency Forest. Proc. East. Snow Conf. (in press).

Thomas, R.B., 1990. Problems in determining the return of a watershed to pretreatment conditions: techniques applied to a study at Caspar Creek, California. Water Resour. Res. 26(9): 2079-2087.

Thomas, R.B., and W.F. Megahan, 1998. Peak flow responses to clear-cutting and roads in small and large basins, western Cascades, Oregon: A second opinion. Water Resour. Res. 34(12): 3393-3403.

Verry, E.S., J.R. Lewis and K.N. Brooks, 1983. Aspen clearcutting increases snowmelt and storm flow peaks in north central Minnesota. Water Resour. Bull. 19(1): 59-67.

Vital A.R.T., W.P. Lima, F. Poggiani and F.R.A.Camargo, 1999a. Biogeoquímica de uma microbacia após o corte raso de uma plantação de eucalipto de 7 anos de idade. Scientia Forestalis, 55: 17-28.

Vital, A.R.T., W.P. Lima and F.R.A.Camargo, 1999b. Efeitos do corte raso de plantação de Eucalyptus sobre o balanço hídrico, a qualidade da água e as perdas de solo e de nutrientes em uma microbacia no Vale do Paraiba, SP. Scientia Forestalis, 55: 5-16.