1. Introduction
The lower surface boundary to the atmosphere moderates roughness, albedo, and emissivity, and can act as a water or heat reservoir, depending on its state. The surface can accordingly modify atmospheric stability, humidity, cloud cover, precipitation, and air temperature, and thus the local and regional surface energy and water balance, through complex interactions.
International studies point to regions where the land surface has a high influence on the local weather and climate, although the strength of the land–atmosphere coupling in southern Norway is still inconclusive. Norway is located in the receiving end of the westerlies that have passed over the North Atlantic. This, combined with a long coast lined with mountains, gives Norway on average nearly 1500 mm of precipitation per year. Evaporation from the land surface is on average less than one-fourth of the received precipitation (Hanssen-Bauer et al. 2009). With a predominantly energy-limited evapotranspiration regime, Norway has not stood out in land–atmosphere studies focusing on soil moisture–atmosphere coupling.
Given Norway’s high latitude and maritime location, representation of sea surface temperature (SST) and snow cover receives considerable attention in regional weather, climate, and hydrological modeling. Lack of observations; computational restraints on model resolution, parameterizations, assimilation routines, and initialization errors; and model biases in interacting variables introduce uncertainty in modeled SST and ground snow. For instance, historical runs (1900–2005) of the CMIP5 models show cold SST biases in the extratropical North Atlantic (Wang et al. 2014). The same models show historical snowfall rates on ice-free land north of 50°N that are nearly twice the amount estimated in the observation-based Water and Global Change (WATCH; Weedon et al. 2011) dataset (Brutel-Vuilmet et al. 2013). Even recent reanalysis datasets such as ERA-Interim may show biases in snow depth (Balsamo et al. 2015).
Recent studies (Wramneby et al. 2010; Rydsaa et al. 2015) suggest that vegetation change deserves more attention in our region. Previous studies comparing the sensitivity of terrestrial atmospheric variables to SST anomalies and land-cover change have found regional (de Noblet-Ducoudré et al. 2012) and seasonal (Findell et al. 2009) impacts of a similar magnitude. While vegetation change is now included in IPCC climate projections, it is usually not included in statistical or dynamical downscaling of these models for our region. On the other hand, impact assessment studies of hydrological changes due to land surface changes are commonly conducted without coupling to the atmosphere.
This study explores the sensitivity of the surface energy and water balance in southern Norway to three surface representations, namely, vegetation, snow, and SST. This is done within the constrained environment of the Advanced Research version of the Weather Research and Forecasting (WRF) Model (ARW) framework (Skamarock et al. 2008) by 1) increasing the boreal forest line, 2) increasing ground snow by altering the snow/rain criterion in the land surface scheme, and 3) perturbing the SST. The experiments are run with a grid spacing of 3.7 km over two hydrological years, the dry 2009/10 and the wet 2010/11. The experiments shed light on the regional importance of each of these surface features within weather and climate models, such as WRF, and on the expected information loss from using stand-alone land surface and hydrological models.
The WRF Model, its model configuration, and the study area are presented in the following section. Section 3 outlines the design of each of the three experimental runs, and the results are presented in section 4. A discussion and conclusions are offered in sections 5 and 6.
2. The WRF Model, initialization, and forcing
The WRF Model framework is widely used around the world. It is used for weather forecasting, research, and regional downscaling in Norway (e.g., Heikkilä et al. 2011) and Europe (e.g., Katragkou et al. 2015). The WRF Model system is a flexible framework, with two dynamical cores. In this study, ARW, version 3.5.1 (Skamarock et al. 2008), is used.
a. The ARW and configuration
The model was configured as listed in Table 1, with the Yonsei University (YSU) planetary boundary layer (PBL) scheme; the Kain–Fritsch cumulus scheme; the Goddard shortwave, longwave, and microphysics schemes; and the MM5 similarity scheme for the surface layer physics. A version of the Noah land surface model (LSM) was used (Mitchell 2001, and references therein), with enhancements to improve snow processes, as described below. Also tested, but rejected, was the option to include a diurnal SST parameterization (Zeng and Beljaars 2005), as it was found to slightly reduce the daily average SST throughout most of the year. A more detailed description of the Noah LSM is given next, with emphasis on vegetation and snow parameterizations.
Model configuration with the outer domain values given in parentheses.
b. The Noah LSM
The unified NCEP–NCAR Noah LSM is an open source, community model with a wide range of users. It is a single-column model with four soil layers, a total soil depth of 2 m, with prognostic soil moisture and temperature and freeze/thaw soil physics. Soil thermal and hydraulic parameters like conductivity, field capacity, and wilting point are available as tabulated values according to local soil type. Lookup tables also give the vegetation root depth, stomatal resistance, and maximum snow-covered albedo according to the grid cell’s dominating vegetation type (Table 2 for parameter values for evergreen needleleaf forest, wooded tundra, and mixed tundra). Noah has a canopy layer storing water, and snow is stored on the ground in a single, bulk layer. The surface temperature is a single, weighted mean of the temperature of snow-covered ground, canopy, and bare soil, according to their gridcell fraction. It has a linearized, noniterative surface energy balance, solving for a slab skin temperature after estimating the evapotranspiration using a modified version of the Penman equation (Mahrt and Ek 1984).
The modified IGBP MODIS Noah vegetation parameters for evergreen needleleaf, wooded tundra, and mixed tundra; that is, the vegetation types that are changed in the Veg experiment. The parameters with a min and max value are scaled according to the spatially and temporally varying green vegetation fraction, as provided by an external, satellite-derived dataset. Parameters 
c. The University of Arizona Noah alterations
The University of Arizona (UA) snow physics for Noah (Wang et al. 2010) are used in this study. The UA alterations aim to redeem a too early snowmelt, and excess snow sublimation and downward sensible heat flux often found in the Noah model (Wang et al. 2010). The physical processes associated with deep snow in boreal forests in spring were specifically considered, including the effect of vegetation shading on snow, increased (under canopy) aerodynamic resistance under stable atmospheric conditions (increasingly so with more vegetation present), and a revised roughness length for snow-covered vegetation.
d. Vegetation data, forcing data, and spinup
The vegetation types are given by the MODIS Boston University IGBP dataset, with alterations made by NCEP to include tundra and lakes. For this study, all vegetation in the open shrubland class was converted to the mixed tundra class, as the original was a 60°N threshold for the conversion that resulted in a spurious schism through the center of southern Norway. Additionally mixed tundra was changed to wooded tundra below 950 m, representing birch forest, to provide better consistency with a national vegetation map (Moen et al. 1998).
The land surface model was spun up, using the NCAR High-Resolution Land Data Assimilation System (HRLDAS; Chen et al. 2007) forced with ERA-Interim data (Dee et al. 2011), cycling the year preceding the experimental runs, 2008/09.
Apart from ground temperature and moisture fields, which were transferred from HRLDAS, the WRF Model was initialized with fields from the ERA-Interim. The model was run in climate mode, initialized on 1 October 2009, and forced with fields from ERA-Interim (updated every 6 h on the outer domains’ lateral boundaries, and every 24 h for SST fields). Spectral nudging was not used in order to allow the model more freedom to react to the changes in surface representation.
e. Study area
This study focuses on South Norway, the southern half of Norway. The integration domains are presented in Fig. 1. The outer domain is the enclosed, outer area of the figure. It consists of 77 × 160 grid boxes, with a grid spacing of about 18.5 km, or 0°10′. The inner domain, shown as a blue rectangle, consists of 156 × 261 grid boxes with a spacing of 3.7 km or 0°02′. In further reference to the study area, only the land area in the inner domain is considered.
The outer area of the figure depicts the outer model integration domain. The inner domain, or study area, is within the blue rectangle. Areas originally covered with evergreen needleleaf are marked with dots. Areas where mixed or wooded tundra are replaced by an evergreen needleleaf forest in the Veg experiment are marked with a dark green color (stars) in the inner (outer) domain.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The outer area of the figure depicts the outer model integration domain. The inner domain, or study area, is within the blue rectangle. Areas originally covered with evergreen needleleaf are marked with dots. Areas where mixed or wooded tundra are replaced by an evergreen needleleaf forest in the Veg experiment are marked with a dark green color (stars) in the inner (outer) domain.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The outer area of the figure depicts the outer model integration domain. The inner domain, or study area, is within the blue rectangle. Areas originally covered with evergreen needleleaf are marked with dots. Areas where mixed or wooded tundra are replaced by an evergreen needleleaf forest in the Veg experiment are marked with a dark green color (stars) in the inner (outer) domain.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
A seasonal snow cover is present in most of Norway. The number of days with snow varies considerably, both spatially and from year to year. Normally, it lasts from 0 to 300 days yr−1, with snow water equivalent (SWE) varying from 0 to 3000 mm (Hanssen-Bauer et al. 2009). Norway receives on average nearly 1500 mm of precipitation per year, with large regional differences, as the Scandinavian mountains split South Norway into a rainy western part and a drier eastern part.
f. Study period
Interannual weather variability in Norway is influenced by the North Atlantic Oscillation (NAO), especially in winter. A negative phase of the NAO is usually concurrent with cold and dry conditions in Norway, while a positive NAO phase usually indicates warm and wet conditions. To evaluate to what degree the sensitivities found vary with weather variability, the study is conducted over a time period when the phase of the NAO changed from positive to negative.
A hydrological year is the 12 consecutive months when the hydrological cycle is on average the most in balance. In the Northern Hemisphere, this is usually from October through September. This study spans two hydrological years: from October 2009 to October 2010, which was a particularly cold and dry hydrological year in Norway with the most negative winter (DJF) NAO on record; and from October 2010 to October 2011, which was a warm and wet year during which the NAO switched from negative to positive midwinter. The mean temperature in 2010 was 1°C below the 1961–90 average, and the precipitation was 85% of the 1961–90 average, which is the fifteenth driest since 1900. The year 2011 was one of the warmest and wettest years in Norway; before 2014 it was the warmest and wettest on record. The mean temperature was 1.9°C above the 1961–90 average, and total precipitation was 125% of the 1961–90 average. The winter of 2011 did, however, start out fairly dry and cold.
g. Model performance
The 2-m temperature 
Relative humidity was compared to observed humidity from more than 100 meteorological stations for 2010/11 using the Model Evaluation Tools (MET) software (Brown et al. 2009) and observational data from NOAA/NWS/NCEP (2008). The model shows a slight positive bias in specific humidity, by approximately 4%, and a slight exaggeration of the diurnal moisture variability in the warm season. The wind speed was evaluated similarly and shows good performance but has slight underestimations of the highest wind speeds.
The modeled snow depth 
The difference between simulated and observed snow cover fraction (Cntrl − obs) in fall (SON), winter (DJF), and spring (MAM).
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The difference between simulated and observed snow cover fraction (Cntrl − obs) in fall (SON), winter (DJF), and spring (MAM).
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The difference between simulated and observed snow cover fraction (Cntrl − obs) in fall (SON), winter (DJF), and spring (MAM).
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
3. Experimental design
After a control run (Cntrl) as described above, three experimental runs were initiated, each perturbing a different surface feature, namely, vegetation, snow, and SST. Table 3 summarizes the three experiments.
The surface perturbation experiments.
a. Increasing the boreal forest line (the Veg experiment)
If global warming continues unabated, the northern extents of boreal forest ecosystems are expected to slowly go through a transition from tundra to open shrubland to woodland to boreal forest (Liess et al. 2012). There is, however, large uncertainty regarding historical, present-day, and future land-cover distribution, and considerably different land-cover inventories and projections are used in weather forecasting and climate model runs (Yucel 2006; Klein Goldewijk et al. 2011; Meiyappan and Jain 2012; Tao et al. 2013; Broxton et al. 2014). Even so, land-use and land-cover (LULC) change were included in CMIP5 because of its importance (e.g., Hurtt et al. 2009; Klein Goldewijk et al. 2011).
Observations by satellites and airplanes and field studies have shown an upward shift in species elevation in Europe (Grytnes et al. 2014) and Norway (Bryn 2008; Wehn et al. 2012). Since 1925, the Norwegian forest volume has tripled (Tomter and Dalen 2014), and in many regions forest lines have expanded to higher altitudes. However, because of historic and current land-use practices, in certain areas the forest line can be as much as 200 m below its present potential (Wehn et al. 2012). Subtle vegetation changes (e.g., vegetation densification or a forest-line advance) are difficult to implement without using a high-resolution model. Feedback effects between warming, snow cover, and vegetation cover may not be adequately addressed in global models unable to resolve the local terrain, and thus temperature gradients, well enough (Giorgi et al. 1997).
To shed light on the sensitivity of the surface energy and water balance to a change in vegetation cover, a simple perturbation of the boreal forest line is implemented in the Veg experimental run. Mixed and barren tundra below 1150 m are replaced with evergreen needleleaf forest, increasing the forest line by about 200 m and changing the vegetation type in about a quarter of the land grid cells (Fig. 1, Table 3).
b. Increasing ground snow (the Snow experiment)
More often than not, models show biases in their snow simulations (Slater et al. 2007; Chen et al. 2014). For instance, in ERA-Interim, forests with snow were given a too low albedo, influencing the timing of spring snowmelt (Balsamo et al. 2015). If such biases significantly affect, for example, modeled precipitation intensity in our region, this needs to be accounted for.
Studies dating back to Yeh et al. (1983) have looked at the effect of ground snow on the following seasons by initializing experimental runs with extra snow. These experiments add moisture to the model. In this study we aim to evaluate and compare the responses of the surface energy and water balance to different changes in surface characteristics. Perturbing the ground snow indirectly by changing the snow/rain criterion for precipitation in the land surface part of the model makes it possible to trace back a response in precipitation to a change in ground snow, rather than having to disentangle it from the effect of extra moisture in the model initialization.
The snow/rain criterion is a feature of the land surface model itself, used to diagnose whether the incoming precipitation should be treated as frozen or liquid (e.g., Yang et al. 1997; Slater et al. 2007; Niu et al. 2011). When Noah LSM is coupled with WRF, by default, the snow/rain criterion relies on the ratio of solid to liquid incoming precipitation as diagnosed in the (atmospheric) microphysical scheme. If this information is not available, the lowest atmospheric model layer temperature with a categorical temperature threshold of 0°C is used.
In the current model configuration, the snow/rain criterion is diagnosed from the ratio of solid to liquid hydrometeors calculated in the Goddard microphysical scheme. Compared to using a categorical temperature threshold of 0°C, the current model setup (using the information from the microphysical scheme) leads to a slightly more snow in coastal areas and slightly less snow in inland areas (not shown).
Snow/rain criteria temperatures in the literature vary from −5° to over 6°C [references within Yang et al. (1997); Wen et al. (2013)]. In the current model configuration, the lowest model layer is at about 27 m above ground; thus, a temperature threshold of 2.5°C is expected to be in the higher end of realistic threshold temperatures for snowfall. In the Snow experiment, a categorical temperature threshold of 2.5°C is used as snow/rain criterion in the land surface model in order to increase ground snow without explicitly adding moisture to the model.
c. The 0.4°C SST perturbation (the SST experiment)
A large part of the variability between the models included in CMIP5 can be attributed to differences in the models’ representation of snow albedo feedback effect and NAO (Cattiaux et al. 2013). Historical runs (1900–2005) of the CMIP5 models show cold SST biases in the extratropical North Atlantic (Wang et al. 2014). Biases in surface observational and/or model SST data may preclude a hindcast, weather forecast, or climate projection. A hit or miss of the internal variation (for instance of the NAO) in an atmosphere–ocean global circulation model (GCM) providing forcing data for a regional model has the potential to deteriorate the results of a regional study (Laprise 2014).
Previous studies have shown that the implementation of SSTs in a regional climate model (RCM) can influence temperature over the European continent (Cattiaux et al. 2011) and Norway (Køltzow et al. 2011). A North Atlantic SST increase of 0.29°C decade−1 for the period 1978–2007 has been observed (Cattiaux et al. 2011). The oceanic warming has been found to have contributed to the European land warming over the same period. Focus in this study is to evaluate the role of feedbacks exerted from changes in the land cover to those from a perturbation in SST. Thus, a uniform SST increase of 0.4°C is implemented in both model domains (Fig. 1). Additional SST sensitivity experiments, such as forcing the model with downscaled SSTs from CMIP5 models, would be of interest but are beyond the scope of this study.
4. Results
Key variables controlling the surface energy and water balance are elaborated on below. The land area of the inner domain (study area) is considered (Fig. 1). Average changes [experimental run (exp) minus control run (Cntrl)] over the study area in ground snow cover, surface temperature 
Mean difference 
Local and seasonal changes were notably larger. Paired sample Student’s t tests, using a 95% confidence interval, were performed at each grid cell to determine whether there are significant changes in the annual and seasonal mean values of the control and experimental runs. Table 5 gives, for each variable and year, the mean change averaged over the area of significant change. Percentage land area where significant changes in annual mean values occurred varied from 7% to 98%. For each variable, where applicable, the largest relative change [(exp − Cntrl)/Cntrl] is also presented.
The area (%) of significant change in mean value from the control run A and the mean difference 
a. Snow cover and energy balance changes
Changes in ground snow cover and the components of the surface energy balance are depicted in Figs. 3–5. Local gridcell changes (exp − Cntrl) in the length of the snow season (the number of days with ground snow) in the two hydrological years are depicted in Fig. 3. Figure 4 depicts 2010/11 monthly change in the surface energy balance induced by each experiment, and Fig. 5 shows the seasonal extent and magnitude of significant changes in mean seasonal 
Change (exp − Cntrl) in the length of the snow season (the number of days with ground snow) in (top) the dry, cold hydrological year 2009/10 and (bottom) the warm, wet hydrological year 2010/11. The study area mean change is denoted in the upper-left corner.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Change (exp − Cntrl) in the length of the snow season (the number of days with ground snow) in (top) the dry, cold hydrological year 2009/10 and (bottom) the warm, wet hydrological year 2010/11. The study area mean change is denoted in the upper-left corner.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Change (exp − Cntrl) in the length of the snow season (the number of days with ground snow) in (top) the dry, cold hydrological year 2009/10 and (bottom) the warm, wet hydrological year 2010/11. The study area mean change is denoted in the upper-left corner.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The 2010/11 study area average monthly change from the control experiment in sensible, latent, and remaining (other) fluxes, for the Veg (first monthly bar), Snow (second monthly bar, marked with diagonal lines), and SST (third monthly bar, marked with horizontal lines) experiments.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The 2010/11 study area average monthly change from the control experiment in sensible, latent, and remaining (other) fluxes, for the Veg (first monthly bar), Snow (second monthly bar, marked with diagonal lines), and SST (third monthly bar, marked with horizontal lines) experiments.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The 2010/11 study area average monthly change from the control experiment in sensible, latent, and remaining (other) fluxes, for the Veg (first monthly bar), Snow (second monthly bar, marked with diagonal lines), and SST (third monthly bar, marked with horizontal lines) experiments.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
(a) Mean 
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
(a) Mean
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
(a) Mean
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
1) Increasing the boreal forest line (the Veg experiment)
The forest-line heightening (the Veg experiment) caused significant changes in ground snow cover fraction in over two-thirds of the study area (Table 5). The average number of days with snow increased by 6 days in 2009/10 and 4 days in 2010/11 (Table 4, Fig. 3), although in certain areas the first snow-free day occurred earlier in 2010/11 (not shown). The experiment caused significant changes in 
Changes in turbulent fluxes caused by the Veg experiment mainly occurred in the warm season (Fig. 4). Where tundra was converted to evergreen needleleaf forest, a shift toward both higher SH and LH can be seen (Table 5); however, certain locations showed a large increase in LH coinciding with an equivalent decrease in SH (not shown). Where the vegetation remained unaltered, some grid cells showed the opposite pattern, that is, a decrease in LH and a comparable increase in SH.
The colder year, 2009/10, showed a slightly larger increase in 
2) Increasing ground snow (the Snow experiment)
The parameterization change applied in the Snow experiment clearly increased ground snow cover. Significant changes occurred in nearly the entire study area (Table 5). The largest increase was found near the coast in spring and fall and at higher altitudes in summer (not shown). Averaged over the two hydrological years, the area with ground snow was increased by around 12%, or 40 000 km2, on 1 November and by around 6%, or 20 000 km2, on 1 May. In the warmer and wetter 2010/11, the mean number of days with ground snow increased by about 2 weeks, whereas in colder and drier 2009/10, it increased by only 1 week (Fig. 3).
The additional snow increased the reflection of shortwave radiation (albedo) and thus decreased 
The SH was significantly reduced in nearly half of the area, by 2–3 W m−2, on average. A small decrease in LH was also found. The largest reduction in turbulent fluxes coincided with the months of maximum insolation and snow cover, that is, April–June (Fig. 4). In most months, slight changes in fluxes other than SH and LH are seen (Fig. 4). This is mostly due to changes in the fluxes associated with melting and freezing on the ground. Annual average 
3) The 0.4°C SST perturbation (the SST experiment)
The SST experiment caused significant changes in ground snow cover in more than 70% of the study area (Table 5). The number of days with ground snow decreased, on average, by 2–3 days (Table 4, Fig. 3). Slight increases in net shortwave radiation were found, but of the three experiments, the SST perturbation had the smallest effect on 
Significant changes in turbulent fluxes were found in less than 21% of the grid cells. The largest changes were found in summer and fall (Fig. 4), when LH increased at the expense of SH. The SST perturbation caused significant changes in annual 
b. Water balance changes
Changes in the components of the water balance are depicted in Figs. 6–8. The extent and magnitude of the 2010/11 significant changes in mean seasonal precipitation and runoff are presented in Figs. 6 and 7. Gridcell evaporation and precipitation changes are presented, for each experiment, in scatterplots in Fig. 8.
As in Fig. 5, but for mean daily P. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
As in Fig. 5, but for mean daily P. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
As in Fig. 5, but for mean daily P. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
As in Fig. 5, but for mean daily R. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
As in Fig. 5, but for mean daily R. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
As in Fig. 5, but for mean daily R. Note that here increases are marked in blue tones, while decreases are marked with red tones.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Scatterplots of the annual mean differences (exp − Cntrl) in E (x axis) and P (y axis). Points originating from areas where the vegetation was changed in the Veg experiment are shown, for reference, for all experiments, with a green, hatched color scheme, with darker colors representing higher densities of points. The remaining points are marked with a black color scheme, with brighter colors representing higher densities of points.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Scatterplots of the annual mean differences (exp − Cntrl) in E (x axis) and P (y axis). Points originating from areas where the vegetation was changed in the Veg experiment are shown, for reference, for all experiments, with a green, hatched color scheme, with darker colors representing higher densities of points. The remaining points are marked with a black color scheme, with brighter colors representing higher densities of points.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Scatterplots of the annual mean differences (exp − Cntrl) in E (x axis) and P (y axis). Points originating from areas where the vegetation was changed in the Veg experiment are shown, for reference, for all experiments, with a green, hatched color scheme, with darker colors representing higher densities of points. The remaining points are marked with a black color scheme, with brighter colors representing higher densities of points.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
The annual mean precipitation can thus be sourced from local evaporation or from advected atmospheric moisture (MFC). Figure 9 shows the annual mean change in MFC for each experiment in 2009/10.
Annual (2009/10) change in MFC from the control run for each experimental run. The mean change over the study area is given in the upper-left corner. Because of the very high variability of the divergence field, a Gaussian filter has been applied.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Annual (2009/10) change in MFC from the control run for each experimental run. The mean change over the study area is given in the upper-left corner. Because of the very high variability of the divergence field, a Gaussian filter has been applied.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
Annual (2009/10) change in MFC from the control run for each experimental run. The mean change over the study area is given in the upper-left corner. Because of the very high variability of the divergence field, a Gaussian filter has been applied.
Citation: Journal of Hydrometeorology 18, 1; 10.1175/JHM-D-15-0146.1
1) Increasing the boreal forest line (the Veg experiment)
The Veg experiment led to significant changes in precipitation in nearly a quarter of the study area (Table 5). The signal of these changes varied with location and season (Fig. 6b). The most significant change in precipitation was found in summer: an increase of 12% in 15% of the study area. The greatest increase in LH was also found in summer (Fig. 4). Generally, where tundra was replaced with forest, an increase in both evaporation and precipitation is evident (Fig. 8). The scatterplot also shows that in some areas where the vegetation remained unaltered, a slight increase in precipitation in combination with a small decrease in evaporation is found. Indeed, comparing Fig. 6b with Fig. 1 shows that some of the areas where significant changes in precipitation occurred are outside the area of vegetation change.
The annual increase in precipitation was nearly as large as the increase in evaporation, averaged over the study area (Table 4). Compared to the control run, a slight moisture flux divergence occurred in the study area (Fig. 9), that is, part of the humidity increase was transported out of the study area. Only a minor decrease in annual runoff was found. Seasonal runoff variability was, however, affected (Fig. 7b), with more runoff in spring and less runoff in summer in western and central Norway. An increase in precipitation is seen in the same areas in spring (Fig. 6b).
2) Increasing ground snow (the Snow experiment)
The Snow experiment only had a minor effect on the annual water balance (Tables 4, 5), shown as a slight decrease in evaporation throughout the year. The greatest change in precipitation was seen in spring in coastal areas, areas that had more snow cover in the Snow experiment than in the control run. In spring, minor decreases in both convective and nonconvective precipitation were found (not shown), with an average reduction of 3% in 13% of the study area, as well as slight reductions in evaporation. In the other seasons, less than 10% of the study area showed significant changes in mean precipitation. Both decreases and increases in precipitation occurred (Fig. 6). In summer, a slight increase in precipitation is seen in inland areas.
The snow cover experiment directly influenced the partitioning of precipitation between rain and snow and induced rather large seasonal shifts in runoff. Coastal areas experienced less runoff in fall and winter, but more in spring. At higher altitudes more water was stored from spring to summer, delaying runoff (Fig. 7c).
3) The 0.4°C SST perturbation (the SST experiment)
On average, the SST experiment intensified the water cycle, with increasing precipitation, evaporation, and runoff (Table 4) in all seasons (Fig. 6d). Average absolute land surface evaporation increased by about one-fifth of the precipitation increase. Most of the precipitation increase was due to an increase in atmospheric moisture flux convergence over the study area (an increase of 0.06 mm day−1; Fig. 9). The increase in MFC was due to an increase in west–east (zonal) moisture transport. Annual runoff increased by nearly 3% averaged over the study area.
The SST experiment induced significant precipitation changes in more than a quarter of the study area. The most extensive change was found in winter, when 35% of the land surface experienced, on average, a 7% increase in precipitation (Fig. 6d). A larger mean relative change (an increase of 9%) in precipitation was found in the summer but in fewer areas. The largest relative change in runoff (an increase of 10%) was seen at the coast in winter (Fig. 7d).
5. Discussion
a. Increasing the boreal forest line (the Veg experiment)
The forest-line heightening, that is, the Veg experiment, caused the largest changes in surface 
The changes in mean temperature, averaged over the study area, were the smallest of the experiments; however, local and seasonal temperature changes were larger than those caused by the SST experiment. Though vegetation shading, as parameterized by the model, delayed the first snow-free day in most areas, snowmelt generally started earlier, and the number of days with a full snow cover on the ground was reduced, allowing the ground to heat up. Warmer 
The effect of the Veg experiment was, for the surface water balance, dominated by an evaporation increase, predominantly found in spring and summer. Annual mean precipitation increased by nearly as much as the evaporation increase. Annual runoff was thus barely affected. Snowmelt generally started earlier, shifting a fraction of runoff from summer to spring. If the results are representative of the sensitivity in other models, it justifies a stronger emphasis on vegetation type inventories, vegetation model development and validation, and increased attention to vegetation change in the region, for example, in regional downscaling of climate models or when using the results from climate models that parameterize vegetation changes. Counteracting precipitation changes may need to be considered when modeling evaporation changes in southern Norway in offline land surface or hydrological models. Changes in vegetation properties may also have affected observed temperature and precipitation trends. Further, the large sensitivity of surface and TOA net radiation found here points to the need for high-resolution studies, including biogeochemical effects, on the impact of vegetation change in our region to provide better information for land management.
b. Increasing ground snow (the Snow experiment)
The Snow experiment increased the number of days in a year with snow cover by 1–2 weeks. Changes in snowfall diagnosis could not occur when the conditions were so cold that the microphysical scheme, which computed the snowfall fraction in the control run, also diagnosed precipitation as snow (section 3). Thus, the impact of the parameterization change was more pronounced in the warmer 2010/11 than in the colder 2009/10, and on the coast in winter and in the high mountains in summer. Because of the additional precipitation diagnosed as snow, an increase in the flux associated with snow melting on the ground was seen. The Snow experiment induced significant changes in 
Of the three experimental runs, the Snow experiment had the smallest effect on annual precipitation and runoff. Seasonal runoff changes were found, with peak runoff increasing and being delayed by more than 1 month. Slight reductions in precipitation were seen, predominantly in late spring and early summer. However, in scattered inland areas, slight precipitation increases were seen. Observational (Kasurinen et al. 2014) and modeling (Rydsaa et al. 2015) studies have shown that evapotranspiration in the boreal regions is primarily controlled by the atmospheric vapor pressure deficit and not by soil moisture. In climates and regions, where the evapotranspiration regime is soil moisture limited, an increase in snow cover may lead to increased soil moisture persisting into the warmer seasons when soil moisture normally would be limited, in turn increasing evaporation and precipitation in these months (e.g., Dutra et al. 2012). This was largely not found here. The results suggest that in the study region and for the current climate, slight differences in modeled snow cover in similar, coupled land surface–atmosphere models may be expected to have little impact on modeled evaporation and precipitation rates.
c. The 0.4°C SST perturbation (the SST experiment)
The SST experiment resulted in the smallest change in 
The annual surface water balance showed a large sensitivity to the rather small SST perturbation. Among the three experimental runs, the SST experiment caused the spatially and temporally most homogeneous changes to the land surface water balance and the largest changes in the annual water balance. Annual terrestrial precipitation rates increased about 5 times more than land evaporation rates, and thus runoff increased. Most of the additional precipitation originated from moisture advected into the study area. The largest changes were seen in winter. These findings, of a stronger sensitivity in terrestrial precipitation and temperature to SST change in the cold seasons, are in line with previous research (Cattiaux et al. 2011; Køltzow et al. 2011). The findings show that rather small biases in SST data may have significant impacts on the modeled, terrestrial water balance and 
d. Limitations
The SST experiment was by design simplistic. By applying a uniform SST increase in both model domains throughout the year, it inflicted perturbations of varying size relative to the mean state of the SST in different regions and seasons. The sensitivity to SST perturbations likely depends on the choice of physical parameterization schemes (Table 1), the size of the model integration domain (e.g., Køltzow et al. 2011), and whether or not nudging is applied. In this study nudging has not been used, as more autonomy was sought to allow land–atmosphere coupling effects to develop. Though the lateral boundary conditions of the model’s rather modestly sized outer domain are expected to constrain the large-scale circulation, the SST perturbation may have caused circulation changes (e.g., Graff and Lacasce 2014). However, this has not been the topic of this study. Similarly, evaluating possible remote or large-scale effects of the vegetation or snow cover change (e.g., Xu and Dirmeyer 2011; Orsolini et al. 2013) has not been considered within the scope of this study.
The Noah LSM is a land surface model of intermediate complexity, without an explicit canopy layer or multiple snow layers. This may impact the robustness of the results, particularly in winter and at night, since correctly modeling the stable boundary layer is generally a challenge for many models (Masson and Seity 2009; Holtslag et al. 2013). Including a prognostic surface boundary layer with an explicit canopy could alleviate some of these issues (Masson and Seity 2009).
The mechanisms involved resulting in precipitation changes due to the vegetation change, and how sensitive this response is to choice of model and model configuration (e.g., Hagos et al. 2014) needs more exploring. Further, a dedicated study on how changing surface characteristics in a particular LSM influence evaporation and runoff rates in an online versus offline model run is an interesting topic for future work.
The current study does not include biogeochemical processes, which influence the effect vegetation change has on climate. Local biogeochemical effects include, for instance, the interaction between cloud cover and surface organic aerosols (SOAs) from vegetation. This effect may reduce the warming caused by the vegetation change (e.g., Scott et al. 2014). Few observational studies have been conducted on SOAs and biogenic volatile organic compounds (BVOCs) in Norway. A recent study (Yttri et al. 2011) estimated the airmass exposure to marine and terrestrial surface types for four Nordic sites. At the Norwegian site, over 90% of the footprint was marine. The site was, however, located less than 25 km from the coast, so a larger airmass exposure to vegetation can be expected farther inland.
6. Summary and conclusions
The sensitivity of a regional climate model to perturbations of three surface features, namely, the vegetation cover, the snow cover, and the sea surface temperature, has been investigated. The surface perturbations reflect both the effect of model initialization and parameterization choices and of surface forcing biases in similar models. High-resolution (3.7 km) runs have been conducted using the WRF Model coupled to the Noah land surface model. The area of investigation is South Norway. Two consecutive hydrological years are considered, 2009/10 (relatively cold and dry) and 2010/11 (relatively warm and wet).
The vegetation was altered in the Veg experiment by increasing the boreal forest-line height by about 200 m, to 1150 m. This led to an increase in annual evaporation in the study area of 8%. Annual runoff was not much affected, as precipitation increased by nearly the same amount. Significant increases in annual net radiation and surface temperature were found. Significant changes were also found outside the area of the vegetation change. The result justifies more emphasis on vegetation model development and validation and attention to vegetation change in the region. The sensitivity also points to the need to consider counteracting precipitation changes when modeling vegetation change in offline land surface or hydrological models.
In the Snow experiment, the ground snow was increased without adding moisture to the model. This was done by altering the snow/rain criterion, that is, the diagnosis of the snowfall ratio of the precipitation, in the LSM. The number of days with snow increased, on average, by 1–2 weeks per year, spring runoff was delayed, and 
In the SST experiment, sea surface temperature was increased with 0.40°C. This remote surface forcing perturbation had the largest effect on runoff of the three experiments. It induced more precipitation over southern Norway, largely due to an increase in moisture transportation from the ocean, only partially reduced by an increase in evaporation. Land surface temperature increased on average by 0.2°C, and significant changes were also found in inland areas. Modest biases in SST forcing data are thus expected to significantly impact weather and runoff projections.
Acknowledgments
We thank the editor and two anonymous reviewers whose suggestions helped improve and clarify this manuscript. HBE was funded by Grant 81077 from the Norwegian Water Resources and Energy Directorate. HBE was also supported by Grant 230616/E10 from the Research Council of Norway. The work forms a contribution to Land–Atmosphere Interactions in Cold Environments (LATICE), which is a strategic research area funded by the Faculty of Mathematics and Natural Sciences at the University of Oslo. The WRF Model runs were performed on the Abel Cluster, owned by the University of Oslo and the Norwegian Metacenter for Computational Science (NOTUR), and operated by the Department for Research Computing at the University of Oslo’s Center for Information Technology [Universitetets senter for informasjonsteknologi (USIT); http://www.hpc.uio.no/].
REFERENCES
Balsamo, G., and Coauthors, 2015: ERA-Interim/Land: A global land surface reanalysis data set. Hydrol. Earth Syst. Sci., 19, 389–407, doi:10.5194/hess-19-389-2015.
Banacos, P. C., and D. M. Schultz, 2005: The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Wea. Forecasting, 20, 351–366, doi:10.1175/WAF858.1.
Barlage, M., X. Zeng, H. Wei, and K. E. Mitchell, 2005: A global 0.05° maximum albedo dataset of snow-covered land based on MODIS observations. Geophys. Res. Lett., 32, L17405, doi:10.1029/2005GL022881.
Bright, R. M., C. Antón-Fernández, R. Astrup, F. Cherubini, M. Kvalevåg, and A. H. Strømman, 2014: Climate change implications of shifting forest management strategy in a boreal forest ecosystem of Norway. Global Change Biol., 20, 607–621, doi:10.1111/gcb.12451.
Brown, B. G., J. H. Gotway, R. Bullock, E. Gilleland, T. Fowler, D. Ahijevych, and T. Jensen, 2009: The Model Evaluation Tools (MET): Community tools for forecast evaluation. 25th Conf. on Int. Interactive Information and Processing Systems (IIPS) for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 9A.6. [Available online at https://ams.confex.com/ams/89annual/webprogram/Paper151349.html.]
Broxton, P. D., X. Zeng, D. Sulla-Menashe, and P. A. Troch, 2014: A global land cover climatology using MODIS data. J. Appl. Meteor. Climatol., 53, 1593–1605, doi:10.1175/JAMC-D-13-0270.1.
Brutel-Vuilmet, C., M. Ménégoz, and G. Krinner, 2013: An analysis of present and future seasonal Northern Hemisphere land snow cover simulated by CMIP5 coupled climate models. Cryosphere, 7, 67–80, doi:10.5194/tc-7-67-2013.
Bryn, A., 2008: Recent forest limit changes in south-east Norway: Effects of climate change or regrowth after abandoned utilisation? Nor. Geogr. Tidsskr., 62, 251–270, doi:10.1080/00291950802517551.
Cattiaux, J., R. Vautard, and P. Yiou, 2011: North-Atlantic SST amplified recent wintertime European land temperature extremes and trends. Climate Dyn., 36, 2113–2128, doi:10.1007/s00382-010-0869-0.
Cattiaux, J., H. Douville, and Y. Peings, 2013: European temperatures in CMIP5: Origins of present-day biases and future uncertainties. Climate Dyn., 41, 2889–2907, doi:10.1007/s00382-013-1731-y.
Chen, F., and Coauthors, 2007: Description and evaluation of the characteristics of the NCAR high-resolution land data assimilation system. J. Appl. Meteor. Climatol., 46, 694–713, doi:10.1175/JAM2463.1.
Chen, F., C. Liu, J. Dudhia, and M. Chen, 2014: A sensitivity study of high-resolution regional climate simulations to three land surface models over the western United States. J. Geophys. Res. Atmos., 119, 7271–7291, doi:10.1002/2014JD021827.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, doi:10.1002/qj.828.
de Noblet-Ducoudré, N., and Coauthors, 2012: Determining robust impacts of land-use-induced land cover changes on surface climate over North America and Eurasia: Results from the first set of LUCID experiments. J. Climate, 25, 3261–3281, doi:10.1175/JCLI-D-11-00338.1.
de Wit, H. A., A. Bryn, A. Hofgaard, J. Karstensen, M. M. Kvalevåg, and G. P. Peters, 2014: Climate warming feedback from mountain birch forest expansion: Reduced albedo dominates carbon uptake. Global Change Biol., 20, 2344–2355, doi:10.1111/gcb.12483.
Dutra, E., P. Viterbo, P. M. A. Miranda, and G. Balsamo, 2012: Complexity of snow schemes in a climate model and its impact on surface energy and hydrology. J. Hydrometeor., 13, 521–538, doi:10.1175/JHM-D-11-072.1.
Dyrrdal, A. V., 2010: An evaluation of Norwegian snow maps: simulation results versus observations. Hydrol. Res., 41, 27–37, doi:10.2166/nh.2010.019.
Findell, K. L., A. J. Pitman, M. H. England, and P. J. Pegion, 2009: Regional and global impacts of land cover change and sea surface temperature anomalies. J. Climate, 22, 3248–3269, doi:10.1175/2008JCLI2580.1.
Giorgi, F., J. W. Hurrell, M. R. Marinucci, and M. Beniston, 1997: Elevation dependency of the surface climate change signal: A model study. J. Climate, 10, 288–296, doi:10.1175/1520-0442(1997)010<0288:EDOTSC>2.0.CO;2.
Graff, L. S., and J. H. Lacasce, 2014: Changes in cyclone characteristics in response to modified SSTs. J. Climate, 27, 4273–4295, doi:10.1175/JCLI-D-13-00353.1.
Grytnes, J.-A., and Coauthors, 2014: Identifying the driving factors behind observed elevational range shifts on European mountains. Global Ecol. Biogeogr., 23, 876–884, doi:10.1111/geb.12170.
Hagos, S., L. R. Leung, Y. Xue, A. Boone, F. de Sales, N. Neupane, M. Huang, and J.-H. Yoon, 2014: Assessment of uncertainties in the response of the African monsoon precipitation to land use change simulated by a regional model. Climate Dyn., 43, 2765–2775, doi:10.1007/s00382-014-2092-x.
Hanssen-Bauer, I., and Coauthors, 2009: Klima i Norge 2100: Bakgrunnsmateriale til NOU Klimatilpasning. Norwegian Climate Centre, 148 pp. [Available online at http://met.no/filestore/klimaiNorge2100_hoyoppl.pdf.]
Heikkilä, U., A. Sandvik, and A. Sorteberg, 2011: Dynamical downscaling of ERA-40 in complex terrain using the WRF regional climate model. Climate Dyn., 37, 1551–1564, doi:10.1007/s00382-010-0928-6.
Holtslag, A. A. M., and Coauthors, 2013: Stable atmospheric boundary layers and diurnal cycles: Challenges for weather and climate models. Bull. Amer. Meteor. Soc., 94, 1691–1706, doi:10.1175/BAMS-D-11-00187.1.
Hurtt, G. C., and Coauthors, 2009: Harmonisation of global land-use scenarios for the period 1500–2100 for IPCC-AR5. iLEAPS Newsletter, No. 7, Integrated Land Ecosystem–Atmosphere Processes Study (iLEAPS), Walingford, United Kingdom, 6–8. [Available online at http://www.ileaps.org/sites/www.ileaps.org/files/iLEAPS_Newsletter_07_June_2009.pdf.]
Kasurinen, V., and Coauthors, 2014: Latent heat exchange in the boreal and arctic biomes. Global Change Biol., 20, 3439–3456, doi:10.1111/gcb.12640.
Katragkou, E., and Coauthors, 2015: Regional climate hindcast simulations within EURO-CORDEX: Evaluation of a WRF multi-physics ensemble. Geosci. Model Dev., 8, 603–618, doi:10.5194/gmd-8-603-2015.
Klein Goldewijk, K., A. Beusen, G. van Drecht, and M. de Vos, 2011: The HYDE 3.1 spatially explicit database of human-induced global land-use change over the past 12,000 years. Global Ecol. Biogeogr., 20, 73–86, doi:10.1111/j.1466-8238.2010.00587.x.
Køltzow, M. A. Ø., T. Iversen, and J. E. Haugen, 2011: The importance of lateral boundaries, surface forcing and choice of domain size for dynamical downscaling of global climate simulations. Atmosphere, 2, 67–95, doi:10.3390/atmos2020067.
Kremsa, J., J. Keček, and E. Kubin, 2015: Comparing the impacts of mature spruce forests and grasslands on snow melt, water resource recharge, and run-off in the northern boreal environment. Int. Soil Water Conserv. Res., 3, 50–56, doi:10.1016/j.iswcr.2015.03.005.
Laprise, R., 2014: Comment on “The added value to global model projections of climate change by dynamical downscaling: A case study over the continental U.S. using the GISS-ModelE2 and WRF models” by Racherla et al. J. Geophys. Res. Atmos., 119, 3877–3881, doi:10.1002/2013JD019945.
Liess, S., P. K. Snyder, and K. J. Harding, 2012: The effects of boreal forest expansion on the summer Arctic frontal zone. Climate Dyn., 38, 1805–1827, doi:10.1007/s00382-011-1064-7.
Lundquist, J. D., S. E. Dickerson-Lange, J. A. Lutz, and N. C. Cristea, 2013: Lower forest density enhances snow retention in regions with warmer winters: A global framework developed from plot-scale observations and modeling. Water Resour. Res., 49, 6356–6370, doi:10.1002/wrcr.20504.
Mahrt, L., and M. Ek, 1984: The influence of atmospheric stability on potential evaporation. J. Climate Appl. Meteor., 23, 222–234, doi:10.1175/1520-0450(1984)023<0222:TIOASO>2.0.CO;2.
Masson, V., and Y. Seity, 2009: Including atmospheric layers in vegetation and urban offline surface schemes. J. Appl. Meteor. Climatol., 48, 1377–1397, doi:10.1175/2009JAMC1866.1.
Meiyappan, P., and A. K. Jain, 2012: Three distinct global estimates of historical land-cover change and land-use conversions for over 200 years. Front. Earth Sci., 6, 122–139, doi:10.1007/s11707-012-0314-2.
Mitchell, K., 2005: The community Noah land-surface model (LSM) user’s guide, version 2.7.1. NOAA/NCEP Doc., 26 pp. [Available online at http://www.ral.ucar.edu/research/land/technology/lsm/noah/Noah_LSM_USERGUIDE_2.7.1.pdf.]
Moen, A., A. Odland, and A. Lillethun, 1998: Vegetasjon. Nasjonalatlas for Norge, Statens Kartverk, Hønefoss, 199 pp.
Mohr, M., 2008: New Routines for Gridding of Temperature and Precipitation Observations for “seNorge. no.” Met. Note 08/2008, Norwegian Meteorological Institute, 40 pp. [Available online at http://met.no/filestore/NewRoutinesforGriddingofTemperature.pdf.]
Myhre, G., O. Boucher, F.-M. Bréon, P. Forster, and D. Shindell, 2015: Declining uncertainty in transient climate response as CO2 forcing dominates future climate change. Nat. Geosci., 8, 181–185, doi:10.1038/ngeo2371.
Niu, G.-Y., and Coauthors, 2011: The community Noah land surface model with multiparameterization options (Noah-MP): 1. Model description and evaluation with local-scale measurements. J. Geophys. Res., 116, D12109, doi:10.1029/2010JD015139.
NOAA/NWS/NCEP, 2008: NCEP ADP global upper air and surface weather observations (PREPBUFR format), May 1997–continuing (updated daily). NCAR Computational and Information Systems Laboratory Research Data Archive, accessed 5 December 2016. [Available online at http://rda.ucar.edu/datasets/ds337.0/.]
Orsolini, Y. J., R. Senan, G. Balsamo, F. J. Doblas-Reyes, F. Vitart, A. Weisheimer, A. Carrasco, and R. E. Benestad, 2013: Impact of snow initialization on sub-seasonal forecasts. Climate Dyn., 41, 1969–1982, doi:10.1007/s00382-013-1782-0.
Pitman, A. J., F. B. Avila, G. Abramowitz, Y. P. Wang, S. J. Phipps, and N. de Noblet-Ducoudré, 2011: Importance of background climate in determining impact of land-cover change on regional climate. Nat. Climate Change, 1, 472–475, doi:10.1038/nclimate1294.
Rydsaa, J. H., F. Stordal, and L. M. Tallaksen, 2015: Sensitivity of the regional European boreal climate to changes in surface properties resulting from structural vegetation perturbations. Biogeosciences, 12, 3071–3087, doi:10.5194/bg-12-3071-2015.
Saloranta, T. M., 2012: Simulating snow maps for Norway: Description and statistical evaluation of the seNorge snow model. Cryosphere, 6, 1323–1337, doi:10.5194/tc-6-1323-2012.
Scott, C. E., and Coauthors, 2014: The direct and indirect radiative effects of biogenic secondary organic aerosol. Atmos. Chem. Phys., 14, 447–470, doi:10.5194/acp-14-447-2014.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., doi:10.5065/D68S4MVH.
Slater, A. G., T. J. Bohn, J. L. McCreight, M. C. Serreze, and D. P. Lettenmaier, 2007: A multimodel simulation of pan-Arctic hydrology. J. Geophys. Res., 112, G04S45, doi:10.1029/2006JG000303.
Tao, Z., J. A. Santanello, M. Chin, S. Zhou, Q. Tan, E. M. Kemp, and C. D. Peters-Lidard, 2013: Effect of land cover on atmospheric processes and air quality over the continental United States a NASA Unified WRF (NU-WRF) Model study. Atmos. Chem. Phys., 13, 6207–6226, doi:10.5194/acp-13-6207-2013.
Tomter, S. M., and L. S. Dalen, 2014: Bærekraftig skogbruk i Norge. Tech. Rep. Norsk institutt for skog og landskap, 241 pp. [Available online at http://www.skogoglandskap.no/filearchive/baerekraftig_skogbruk_web.pdf.]
Wang, C., L. Zhang, S.-K. Lee, L. Wu, and C. R. Mechoso, 2014: A global perspective on CMIP5 climate model biases. Nat. Climate Change, 4, 201–205, doi:10.1038/nclimate2118.
Wang, Z., X. Zeng, and M. Decker, 2010: Improving snow processes in the Noah land model. J. Geophys. Res., 115, D20108, doi:10.1029/2009JD013761.
Weedon, G. P., and Coauthors, 2011: Creation of the WATCH forcing data and its use to assess global and regional reference crop evaporation over land during the twentieth century. J. Hydrometeor., 12, 823–848, doi:10.1175/2011JHM1369.1.
Wehn, S., G. Olsson, and S. Hanssen, 2012: Forest line changes after 1960 in a Norwegian mountain region implications for the future. Nor. Geogr. Tidsskr., 66, 2–10, doi:10.1080/00291951.2011.644320.
Wen, L., N. Nagabhatla, S. Lü, and S. Y. Wang, 2013: Impact of rain snow threshold temperature on snow depth simulation in land surface and regional atmospheric models. Adv. Atmos. Sci., 30, 1449–1460, doi:10.1007/s00376-012-2192-7.
Wramneby, A., B. Smith, and P. Samuelsson, 2010: Hot spots of vegetation–climate feedbacks under future greenhouse forcing in Europe. J. Geophys. Res., 115, D21119, doi:10.1029/2010JD014307.
Xu, L., and P. Dirmeyer, 2011: Snow–atmosphere coupling strength in a global atmospheric model. Geophys. Res. Lett., 38, L13401, doi:10.1029/2011GL048049.
Yang, Z.-L., R. E. Dickinson, A. Robock, and K. Y. Vinnikov, 1997: Validation of the snow submodel of the Biosphere–Atmosphere Transfer Scheme with Russian snow cover and meteorological observational data. J. Climate, 10, 353–373, doi:10.1175/1520-0442(1997)010<0353:VOTSSO>2.0.CO;2.
Yeh, T.-C., R. T. Wetherald, and S. Manabe, 1983: A model study of the short-term climatic and hydrologic effects of sudden snow-cover removal. Mon. Wea. Rev., 111, 1013–1024, doi:10.1175/1520-0493(1983)111<1013:AMSOTS>2.0.CO;2.
Yttri, K. E., and Coauthors, 2011: Source apportionment of the summer time carbonaceous aerosol at Nordic rural background sites. Atmos. Chem. Phys., 11, 13 339–13 357, doi:10.5194/acp-11-13339-2011.
Yucel, I., 2006: Effects of implementing MODIS land cover and albedo in MM5 at two contrasting U.S. regions. J. Hydrometeor., 7, 1043–1060, doi:10.1175/JHM536.1.
Zeng, X., and A. Beljaars, 2005: A prognostic scheme of sea surface skin temperature for modeling and data assimilation. Geophys. Res. Lett., 32, L14605, doi:10.1029/2005GL023030.
Zhang, M., and Coauthors, 2014: Response of surface air temperature to small-scale land clearing across latitudes. Environ. Res. Lett., 9, 034002, doi:10.1088/1748-9326/9/3/034002.
