Soil moisture simulations on a regional level : the ability of groundwater model MIPWA to replicate soil moisture observations in Twente

Gurp, H. van (2016) Soil moisture simulations on a regional level : the ability of groundwater model MIPWA to replicate soil moisture observations in Twente.

Abstract:The groundwater model MIPWA simulates the groundwater levels in the North-Eastern part of the Netherlands. For this purpose MIPWA utilizes the unsaturated zone model MetaSWAP. The simulations of this unsaturated zone model have only been verified (van Walsum & Veldhuizen, 2011), no calibration or validation has been performed. Research by Mehrjardi (2015) and Schuurman et al. (2011) suggested that the simulated soil moisture content by MetaSWAP can be improved. The objective of this research is to evaluate and potentially improve the ability of MIPWA to simulate the soil moisture by comparing the simulations of MIPWA to field measurements. This research compares the MIPWA model results to measurement of the ITC soil moisture monitoring network. At 20 observation sites the soil moisture content has been measured at various depths by the ITC soil moisture network. From these measurements, characteristics have been derived which the model should replicate. The measurement data has been explored to identify potential errors, which have been removed during the measurement data preparation. The measurements showed that the measured porosity differs from the expected porosity based on soil type. At 9 observation sites porosities have been measured deviate significantly from the soil type based porosity. This is partly due to disturbances in the soil, as the recorded soil moisture content could not have been recorded in the undisturbed soil at 4 observation sites. The soil moisture content over time has a similar pattern to the evapotranspiration and precipitation deficit. As a result the soil moisture content in the spring and summer lower than the rest of the year. The correlation between meteorological condition and the soil moisture content furthermore showed different behaviour in different layers in the soil. The probes near the surface are more sensitive to precipitation than deeper located probes, while the deeper located probes are more similar to the trend of evapotranspiration and precipitation deficit. As a result the probes near the surface measure large variance in soil moisture content, which dampens with increasing depth. Two classifications of the observation sites have been made. The first classification divides the observations sites based on the groundwater level. Observation sites with a groundwater table close to the surface have higher soil moisture content than the sites with a deep groundwater table. This difference in soil moisture content in visible throughout the year, expect during dry period. The second classification is based on soil type. Two soil types are dominant on the surface of Twente, loamy sandy soils and sandy soils. At the surface the sandy soils contain less moisture than the loamy sandy soil, as was expected based on the soil properties. The MIPWA model simulation period has been extended from 2001 to 2012 by extending the meteorological dataset of the model. This allows the model to be compared to the measurements between 2010 and 2012. The comparison shows that the model is able to simulate the dynamics of soil moisture content in the root zone. The model is able to explain on average 71% of the variance in the soil moisture content in the root zone, but the model does systematically underestimate the soil moisture content. The individual observation sites that perform best are sites with similar measured and simulated porosity, a groundwater table far underneath the surface and have a loamy sandy soil type. In general, observation sites with a similar measured and simulated porosity are able to explain more of the model variance and are able to simulate the soil moisture content with a smaller root mean square error. The smaller error is also present for observation sites with a groundwater table located further away from the surface. The loamy sandy soils are in general better simulating soil moisture dynamics than sandy soils. The other observation sites are quite able to replicate the soil moisture content in the root zone, however can be improved. The observation sites which required improvement are observation sites with a groundwater table close to the surface. The soil moisture stress at these sites is not well simulated as Schuurman et al. (2011) already indicated. The lack of soil moisture stress increases evapotranspiration rates. This decreases the groundwater level in regions with a high groundwater table. The capillary rise has been limit, which allowed the model to simulate the soil moisture stress better and increase the groundwater tables. To improve the soil moisture simulations the modelled porosity has been increased to better match the observed soil moisture content. This significantly reduced the root mean square error of the modelled soil moisture content. The combination of the measures proved successful in improving both soil moisture dynamics and absolute value.
Item Type:Essay (Master)
Faculty:ET: Engineering Technology
Subject:56 civil engineering
Programme:Civil Engineering and Management MSc (60026)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page