Forest Investment Account (FIA) - Forest Science Program
FIA Project M075039

    Peak Flow and Water Yield Responses to Mountain Pine Beetle Infested and Salvage Logged Watersheds [2006/2007 project description only]
Project lead: Alila, Younes
Imprint: Vancouver, BC : University of British Columbia, 2007
Subject: Forest Investment Account (FIA), Dendroctonus Ponderosae, British Columbia
Series: Forest Investment Account (FIA) - Forest Science Program
Lodgepole pine is an abundant species in the interior of British Columbia (BC). The whole province of BC has approximately 15 million hectares of this species, and it is economically important to the province. Pine supplies as much as 80% of the annual timber harvest in some central interior forest districts and comprises 25% of the province’s timber supply (FPB 2004). Mountain Pine Beetle (MPB) infestations are natural forest landscape disturbances. Historically, colder winters in BC are believed to be the natural line of defense against MPB. An early fall cold snap of -30 C for as little as one day will kill pine beetles. Climate change raising temperature through western Canada has caused the beetle to spread out of control. For example, 1985, when the Vanderhoof Area of the central interior of BC was last hit with – 37 C Halloween, was the last time pine beetles in BC met a killing cold in the fall. By the summer of 2005, there were 10 million hectares of infested pine in BC (FPB 2004). According to Hamilton (2005), the MPB is expected to kill 80% of the 1.2 Billion cubic meters of BC pine by 2013. The loss of this much timber would cost an estimate of $2 Billion per year to government coffers for years to come. The MPB infestations have now moved so far south that cold winters are not likely effectively able to control the spread. MPB has made its way over to the Rockies from both northern and southern BC. Experts are warning that it will make its way across the boreal forest and south into the US. Disaster in BC could soon be a national and possibly a continental epidemic particularly if the beetles start infesting the contiguous areas of Jack Pine east of the Rockies, in the Boreal Forest and South of the US-Canada borders (Hamilton 2005). As part of the efforts to controlling the spread of the infestations and minimizing the economic losses, the BC government is adapting a strategy for the salvage logging of MPB affected areas. However, the beetles are killing trees much faster than the BC Forest Industry can harvest them! The primary focus of this project is on the effects of the mountain pine beetle infestations and salvage logging on streamflow characteristics (peak flows, water yield, timing of the flows, etc.). Streamflow is an essential variable in understanding the functioning of watersheds and associated ecosystems because it supplies the primary medium and source of energy for the movement of water, sediment, organic material, nutrients, and thermal energy (Ziemer and Lisle 1998). The mountain pine beetle epidemic and subsequent salvage logging is creating disturbances over larger watersheds (hundreds of square km) with unprecedented high values of equivalent cut area (ECA). The effects of these disturbances on the watershed streamflow response could be dramatic and can not be inferred from the current literature. In BC and elsewhere in Canada, research work on the relationship between forest stand characteristics and hydrology has been conducted at the stand level (Winkler et al. 2005, Pomeroy et al. 2002, Pomeroy and Dion 1996, Pomeroy et al. 1998, Hedstrom and Pomeroy 1998, McNay et al. 1988). How such knowledge translates to the watershed scale remains an elusive question. Internationally, our present knowledge of streamflow changes resulting from forest disturbances is based on paired watershed studies conducted on small drainages that are often less than few square km in size. Concerns have been raised about the validity of extrapolating experimental results from these small paired watershed studies to larger watersheds because hydrologic processes and relationships of these types of watersheds are expected to be different (Benda and Miller, 2001). Only few snow-dominated paired watershed studies over larger drainages have been reported in the literature (Cheng 1989, Potts 1984, Bethlahmy 1975, Burton 1997). However, the size of these watersheds and the level of disturbances in these watersheds are not large enough to be of relevance to the questions raised in this study. Therefore, there is a scientific and practical need for studies that provide some understanding of large scale forest disturbances on watershed scale functions such as streamflows and hydrology as whole. One of the fundamental conditions in paired watershed studies is that the disturbed (treatment) and undisturbed (control) catchments have to be similar in climate and physiography. This condition can only be achieved over small catchments and therefore a paired watershed approach is not even an option for studying forest disturbances over larger catchments. A more suitable (and perhaps the only) method available for the study of disturbances over larger watersheds is to synergistically supplement experimental results with modelling (Ziemer et al. 1991). A hydrologic model (i) alleviate problems associated with paired watershed studies by acting as a control to filter out the effects of climate variability, (ii) allows same watershed to act as its own control, and (iii) is useful for integrating stand-level information into a physically or conceptually reasonable description of basin response to forest disturbances. The capabilities of existing hydrologic models, however, are often overstated and very rarely critiqued. The quantitative results of these models are highly questionable because of uncertainties in their predictions. For instance, the literature is full of hydrologic models that have been calibrated to produce the flows measured at the outlet of a watershed. At the same time, model predictions of detailed flow characteristics at the tributaries and of other internal above and below surface hydrologic processes within the same watershed are rarely verified against measurements. Under these conditions, model predictions of these internal characteristics and processes could be purely an artifact of the model and input parameters and cannot be assumed to be estimates of reality. Until the internal accuracy of models has been established, it is not possible to use model estimates of water quantity and quality for any predictive purposes such as assessing the impact of land use change and forest management on these processes (Grayson et al. 1992). Based on these premises, we collaborated over the last five years with our colleagues from the BC Ministry of Forest on the use of hydrologic models to quantify the effects of forest management on hydrology in several experimental watersheds following an approach outlined in Alila and Beckers (2001). This approach doesn’t require the expensive long-term data collection exercise of a paired watershed study but uses a shorter period of extensively monitored hydro-climate data to develop, test, and calibrate watershed model applications and use them subsequently in a long-term simulation experiment to generate time series of watershed hydrologic response for alternative forest disturbance scenarios. A longer term proxy climate data generated to reflect the regional and at-site climatic characteristics of the study watershed is used as input to the watershed models (Schnorbus and Alila 2004a). We have successfully applied this new approach using the quasi physically based Distributed Hydrology Soil Vegetation Model (DHSVM) in the smaller heavily instrumented Penticton, Redfish, and Carnation Creek experimental watersheds (5 to 25 square km) to quantify the long-term effects of different forest management scenarios on streamflow characteristics and published the work in highly ranked science journals (Whitaker et al. 2002 and 2003, Thyer et al. 2004, Schnorbus and Alila 2004b, Beckers and Alila 2004). The innovative aspect in these publications is the fact that models were developed, parameterized, tested, and calibrated using not only streamflows at the outlet of a watershed but also many other monitored internal above and below surface catchment processes. This project is an extension of the same line of work but to larger watersheds (100s of square km) using a suite of ranging complexity hydrological models that we selected specifically to investigate different types of MPB hydrology questions. There are two types of hydrologic models. First, there is the fully distributed quasi-physically based models such as the DHSVM of Wigmosta et al. (1994), designed for heavily instrumented forest watersheds. As mentioned earlier, our research group has extensive experience in the use of this DHSVM model at heavily instrumented smaller watersheds. However, this type of model with heavy input data requirements is not useful for routine operational purposes at larger ungauged watersheds or at watersheds with much less available data. Second, there are the lumped to semi-distributed conceptual models such as the UBCWM of Quick (1995) or HBV of Lindström et al. (1997). The UBCWM has been developed by one of our hydrologists Prof. M. Quick for BC hydrologic, climatic, and physiographic regimes. It was designed for very large basins with much less input data. As a result, the UBCWM has been and continues to be extensively used by BCMoE, BCHYDRO, ALCAN and other private consultants. Therefore, UBCWM is an operational model that had long been accepted by government and industry in BC as a decision support tool. However, until recently such operational model had never been tested for forestry applications. This last fiscal year, I have been contracted by the BC MoF to investigate the possible use of the UBCWM for predicting streamflows of MPB infested and log salvaged watersheds. Preliminary results indicate that the UBCWM provides good potentials for forestry applications. However, due to the lack of available hydro-climate data at the appropriate basin scales, the model needs to be further tested and better parameterized before it can be adapted to forestry applications.
Contact: Alila, Younes, (604) 822-6058,

Updated August 16, 2010 

Search for other  FIA reports or other Ministry of Forests and Range publications.

Please direct questions or comments regarding publications to