Impacts of topography and land use changes on the air surface temperature and precipitation over the central Peruvian Andes
Introduction
It is well known that topography and land use cover data are key to the performance of regional climate models (RCMs). Topography is particularly important in cloud formation and the distribution of temperature with altitude via atmospheric dynamics such as local and planetary boundary layer (PBL) processes. On the other hand, land use cover data play an important role in the distribution of the energy balance of the surface via variables, such as albedo, roughness height and emissivity, and the local to regional atmospheric circulation, among other factors. In this respect, various studies have shown that improving the representation of these features by using updated land use data and high-resolution topography is important to improve the performance of the Weather Research and Forecasting (WRF) model outputs (e.g., Pitman et al., 2004; Grossman-Clarke et al., 2005; Grossman-Clarke et al., 2010; Lee and Berbery, 2012; Cheng et al., 2013; De Meij and Vinuesa, 2014; Teixeira et al., 2014; Schicker et al., 2016; Jiménez-Esteve et al., 2018). In particular, improving the modeling of air surface temperature and precipitation at high resolution is especially useful in complex mountain regions such as the tropical Andes as forcing inputs to hydrological and glaciological models (e.g., Mourre et al., 2016; Heredia et al., 2018).
The Peruvian central Andes (PCA) is characterized by a very complex climate, with a strong zonal gradient from the west dry region close to the Pacific Ocean toward the humid Amazonian plains to the east (e.g., Garreaud, 1999; Killeen et al., 2007), including a maximum rainfall zone in the eastern slope of the Andes, also called “rainfall hotspot”, where annual precipitation reaches approximately 6000 mm∙year−1 (Espinoza et al., 2015; Chavez and Takahashi, 2017). Previous studies have used the WRF model at high resolution in the PCA region and have identified strong WRF precipitation biases (up to 300%) using different parameterization schemes (e.g., Mourre et al., 2016; Junquas et al., 2018). The major model precipitation biases were generally found over the eastern slope of the Andes. Studies that analyzed surface temperature biases in WRF simulations are scarce in the PCA region and need to be addressed. In the Ecuadorian Andes, Ochoa et al. (2016) found strong biases (up to 8 °C). These works have used the same datasets of topography and land cover from the United States Geological Survey (USGS), which are default datasets in the WRF model. Therefore, in this study, we aim to evaluate the WRF model performance in the PCA region in terms of precipitation and air surface temperature by using a more up-to-date land use cover developed by Eva et al. (2004), and by using a different Digital Elevation Model (DEM) with 30 arc-second spatial resolution (or approximately 1 km) from the Shuttle Radar Topography Mission (SRTM30 product; Farr et al., 2007).
As previously mentioned, several studies have shown that modeling can be sensitive to the representation of both topography and land use cover datasets. Some authors have performed experiments with the WRF model using different datasets. For example, Teixeira et al. (2014) modeled precipitation using the SRTM as topography forcing. They showed differences of up to 500 m between USGS topography data and SRTM in the Portuguese Madeira Island. They also found changes in precipitation patterns associated with the change in topographic forcing, with increases in the mountains and decreases in the valleys. Cheng et al. (2013) found enhancements in air temperature using a land use dataset built from SPOT (Satellites Pour l'Observation de la Terre) satellite images in the Taiwan area. This SPOT product gave better results than the two land use datasets provided by WRF, USGS and a product obtained from MODIS (Moderate Resolution Imaging Spectroradiometer). Using the European CORINE (Coordination of Information on the Environment) land cover, De Meij and Vinuesa (2014) showed improvements in the simulation of the temperature and air quality by changing the land use in zones that were converted to urban zones. Grossman-Clarke et al. (2005) stated that changes in land use affect the turbulent fluxes and planetary boundary layer (PBL) height. In addition, Kim et al. (2013) found that the impact of using updated land cover could show a larger impact on the simulated meteorological variables than changing the parameterization of PBL schemes. Other studies conducted by Schicker et al. (2016), Grossman-Clarke et al. (2010), Lee and Berbery (2012) and Pitman et al. (2004) also showed changes in temperature or precipitation when modifications of land use were taken into account.
Studies using high resolution WRF modeling over the Andes, as mentioned before, have not considered changes of topography and land use cover and how these changes could affect the WRF results. In this work, we evaluate these changes and how they could impact the local atmospheric circulation, precipitation and air surface temperature during the January months from 2004 to 2008, corresponding to the core of the wet season in the PCA region (e.g., Segura et al., 2019). Our study area covers the PCA region from 10°S-14.25°S in latitude and 78°W-73°W in longitude. This area also covers the Mantaro Basin (hereafter called the “MB”), one of the largest basins in the Peruvian Andes (between 500 and 5500 masl) and one whose agricultural activity provides products to Lima, the 9.5 million-population capital of Peru. The study area also includes the coastal front of the Pacific Ocean and the Andes-Amazon transition region, which are located west and east of the Andes Mountains, respectively.
This study is organized as follows. Data and methods are discussed in Section 2. Section 3 covers the WRF simulations with updated land use and topography datasets. This section also evaluates the performance of the experiments that considered the observed daily data of rainfall and minimum and maximum temperature from in situ meteorological stations. Then, a general discussion is given in Section 4. Finally, in Section 5, conclusions are presented based on the results and some suggestions are made about future atmospheric simulations for the Andes.
Section snippets
Data and methodology
Fig. 1a and b show the study region and its location in the western border of tropical South America, respectively. In addition to the MB, Fig. 1a also shows three additional regions analyzed in this work: the Along the Coast (AC) region, which is separated by the 2000 masl contour line from the High Western Slope of the Andes (HWSA), which extends to the western border of the MB. The region from the eastern border of the MB to the lowland Amazon (Andes-Amazon transition region) is called the
Surface parameters
LAI in CTRL or SIM01 (Fig. 4a) increases eastward with the appearance of vegetation from the barren or sparsely vegetated category in the AC to the forest categories in the ESA. There is a predominant increase of LAI when comparing SIM02 with SIM01 due to the appearance of shrubland/grassland in the AC, forest approximately 3500 masl in the HWSA, cropland/grassland in the MB (especially in the Central Mantaro Valley), and forest in the ESA. In this respect, changes in LAI are significant in
Impacts on rainfall
It was found that daily rainfall outputs of SIM01 and SIM02, or those simulations that consider SRTM topography, showed fewer differences with the observed rainfall values. This is illustrated by the improvement of statistics skills. In other words, the improvement was mainly produced by considering the SRTM database. These findings were obtained using 57 meteorological stations and the nearest grid points in the simulations (CTRL, SIM01 and SIM02). However, important changes were observed in
Concluding remarks and recommendations
In the present study, high spatial resolution simulations were conducted using the WRF model version 3.7.1 by changing its topography database from hgt_USGS (default) to hgt_SRTM (implemented) and by changing the land use from lu_USGS (default) to lu_ENEW (implemented). The changes of the precipitation, the air surface temperature and the local atmospheric circulation were evaluated according to changes of the topography and land use databases. The observed data of meteorological stations were
Declaration of Competing Interest
The authors declare that they have no conflict of interest.
Acknowledgments
This research was supported by the project “Study of the physical processes that control the superficial fluxes of energy and water for the modeling of frosts, intense rains and evapotranspiration in the central Andes of Peru” funded by INNÓVATE-PERÚ, Peru, (contract 400-PNICP-PIBA-2014). Jhan-Carlo Espinoza and Clementine Junquas were partially supported by the French AMANECER-MOPGA project funded by ANR and IRD, France (ref. ANR-18-MPGA-0008). Simulations were possible thanks to the
References (53)
- et al.
Land cover change effects on the climate of the La Plata Basin
J. Hydrometeorol.
(2012) - et al.
Vertically integrated moisture flux convergence as a predictor of thunderstorms
Atmos. Res.
(2007) - et al.
Elevation-dependent warming of land surface temperatures in the Andes assessed using MODIS LST time series (2000–2017)
Int. J. Appl. Earth Obs. Geoinf.
(2019) - et al.
Transpiration and evapotranspiration from maize as related to leaf area index
Agric. For. Meteorol.
(1989) - et al.
The use of moisture flux convergence in forecasting convective initiation: historical and operational perspectives
Weather Forecast.
(2005) - et al.
Orographic rainfall hot spots in the Andes-Amazon transition according to the TRMM precipitation radar and in situ data
J. Geophys. Res.
(2017) - et al.
Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part I: model implementation and sensitivity
Mon. Weather Rev.
(2001) - et al.
Investigation of the effects of different land use and land cover patterns on mesoscale meteorological simulations in the Taiwan area
J. Appl. Meteorol. Climatol.
(2013) - et al.
Relations between evaporation coefficients and vegetation indices studied by model simulations
Remote Sens. Environ.
(1994) - et al.
Effects of clouds, soil moisture, precipitation, and water vapor on diurnal temperature range
J. Clim.
(1999)