| Abstract |
Water scarcity has become an issue world-wide, especially in arid and semi-arid environments (or, water-limited environments – WLE) which occupy almost half of the global land area. More than ninety percent of the people living in these environments depend on groundwater for domestic and livestock water supply whereas the omnipresent farming practice is rain-fed agriculture. Despite this, the underlying water flow processes along the soil-vegetation-atmosphere interface in WLE are scientifically not yet well understood complicating a proper management of groundwater resources and agricultural planning. In order to remedy this situation, interdisciplinary approaches are required.
The main objective of this PhD is to contribute to the development of practice-oriented approaches for the investigation of ecohydrological processes along the soil-vegetation-atmosphere interface with a focus on WLE. In view of the complexity of the system soil-vegetation-atmosphere, specific research aspects were selected: i) rainfall characteristics and their impact on rain-fed agriculture; ii) the determination of groundwater recharge; and iii) the quantification of rooting depths and water vapor transport processes in thick unsaturated zones. First and foremost, this PhD thesis aims at developing straightforward, yet scientifically accurate methods which enable quantitative estimates rather than purely qualitative descriptions. Secondly, the proposed approaches are meant to be appropriate for their application in data scarce environments and support the parameterization of unsaturated zone models. Finally, this thesis aims to contribute to the improvement of the current understanding of the complex system soil-vegetation-atmosphere in WLE.
The methods used were manifold and chosen according to the availability of data as well as the spatiotemporal scale of the particular research aspect. For the investigation of rainfall characteristics and their impact on rain-fed agriculture (study i), a research framework composed of several methods was developed. These methods constitute a multi-criteria analysis of rainfall characteristics, fitting of the soil-vegetation-atmosphere transfer (SVAT) model DAISY and subsequent uni- and multivariate analyses between rainfall characteristics and agricultural yields using the artificial neural network approach of self-organizing maps (SOM). Study i) is solely based on remote sensing data and thus could be applied on a large spatial scale and multiple (12) years. In contrast, studies ii) and iii) are field investigations at the point scale and the data used was completely self-collected. Hence, these studies were rather short-term. The main methodologies used in these studies were specifically designed on-site experiments utilizing stable water isotopes as artificial tracer. In study ii), the well-established peak displacement method for the estimation of groundwater recharge was adapted and improved for dry environments using deuterated water (2H2O) as tracer. In the last investigation (study iii) it is made use of the same tracer but with a depth-controlled labeling strategy in order to investigate rooting depths, trace water uptake by plants and monitor the distribution of 2H2O (i.e. water vapor transport) over one dry season. After sample collection and cryogenic vacuum-extraction of soil and plant samples the resulting water samples were subsequently analyzed for their isotopic concentrations using cavity ring-down spectroscopy (CRDS). For the interpretation of this data dual-isotope plots as well as isotopic depth-profiles were evaluated (studies ii and iii).
The main outcome of this thesis is a contribution to an improved understanding of processes governing water transport at the soil-vegetation-atmosphere interface in WLE. In particular, the results from the presented studies highlight the importance of the intra-seasonal distribution of rainfall for agricultural success (study i), provide methodologies to illuminate water transport processes in soils and plants and reveal the need to consider transport processes (i.e. through water vapor and deep tapping roots) in thick unsaturated zones (studies ii and iii). The methodologies developed provide scientifically precise, yet straightforward, practice-oriented approaches that can be applied with minimum prior knowledge even in data-scarce regions.
Study i) reveals the importance of the intra-seasonal distribution of rainfall. Using total rainfall amounts, total counts of dry/wet or wet spells days as single indicators for evaluating or predicting yields from rain-fed agriculture is not sufficient. Rather than that results show that indicators related to durations (rainy season duration, maximum dry/wet spell duration) are more suitable proxies for agricultural success. Inconsistent rainy seasons, i.e. many shifts of wet and dry spells, were found to affect yields most negative. The multivariate analysis using SOM further reveals that total duration of the rainy season as indicator conjunctively with total rainfall, maximum wet and dry spell durations and the number of dry spells are main determinants for current success of rain-fed agriculture. These findings have implications for future agricultural planning. For instance, yields could improve drastically, if the maximum dry spell could be bridged by a small-scale irrigation. The developed modeling framework can be applied for management analyses or studies on the impact of future climatic changes and easily be adapted if better data becomes available.
The use of water stable isotopes as artificial tracer (studies ii and iii) was proven to yield quantitative estimates of groundwater recharge and valuable additional information for understanding water vapor transport mechanisms and the parameterization of SVAT models. One major outcome of study ii) is that using 2H2O as artificial tracer is appropriate for monitoring water transport in dry environments over longer (> 3 months) time scales. Hence, this study and subsequent improvements of the peak-displacement method provide an additional method for the portfolio of techniques to estimate groundwater recharge in WLE. The depth-controlled application of 2H2O (study iii) resulted in a novel method for investigating rooting and root water uptake depths. An additional outcome of this study is the potential to illuminate water vapor movement, which might have a major impact on the sub-surface water balance of WLE.
|
? |
| Citation |
Beyer, M. (2016): Quantitative studies along the soil – vegetation – atmosphere interface of water – limited environments: practice-oriented approaches based on stable water isotopes, modeling and multivariate analysis. Mitteilungen des Instituts für Wasserwirtschaft, Hydrologie und landwirtschaftlichen Wasserbau, Leibniz Universität Hannover, Heft 102, ISSN 0343-8090 |
? |
