THE IMPACT of RAINFALL and NON-RAINFALL on SOIL MOISTURE DYNAMICS in the NAMIB DESERT Bonan Li Submitted to the Faculty of the U

THE IMPACT OF RAINFALL AND NON-RAINFALL ON SOIL MOISTURE

DYNAMICS IN THE NAMIB DESERT

Bonan Li

Submitted to the faculty of the University Graduate School in partial fulfillment of the requirements for the degree Master of Science in the Department of Earth Sciences, Indiana University

August 2017

Accepted by the Graduate Faculty, Indiana University, in partial

fulfillment of the requirements for the degree of Master of Science.

Master’s Thesis Committee

______Lixin Wang, Ph. D.,Chair

______Lin Li, Ph .D.

______Pierre-André Jacinthe, Ph.D.

ⅱ

Acknowledgements

I would like to thank my primary advisor, Dr. Lixin Wang, for his instruction, dedication, patience and all the support he has given to me in the course of my degree pursuit. I also would like to thank my co-advisor Dr. Lin Li and my committee member Dr.

Pierre-André Jacinthe for their insightful comment and encouragement during the last two years at IUPUI.

I thank Farai for his dedication on the field data collection and the data sets were instrumental for my master thesis. My sincere thank goes to all the graduates, undergraduates and staff members in the Department of Earth Sciences. I specially thank

Igor who is one my best friends and colleagues for his generous help in everyday life and academia.

Last, I would like to thank my family members for being my biggest support over the years. Without their encouragement and support, I would not complete my master research.

iii

ABSTRACT Bonan Li

THE IMPACT OF RAINFALL AND FOG ON SOIL MOISTURE DYNAMICS IN THE

NAMIB DESERT

Soil moisture is a key variable in dryland ecosystems. Knowing how and to what extent soil moisture is influenced by rainfall and non-rainfall waters (e.g., dew, fog, and water vapor) is essential to understand dryland dynamics. The hyper-arid environment of the Namib Desert with its frequent occurrence of fog events provides an ideal place to conduct research on the rainfall and non-rainfall effects on soil moisture dynamics. Rainfall and soil moisture records was collected from three locations (gravel plain at Gobabeb

(GPG), sand dune at Gobabeb (SDG), and gravel plain at Kleinberg (GPK)) within the

Namib Desert using CS655 Water Content Reflectometer and tipping-buckets, respectively.

The fog data was collected from the FogNet stations. Field observations of rainfall and soil moisture from three study sites suggested that soil moisture dynamics follow rainfall patterns at two gravel plain sites, whereas no significant relationships was observed at the sand dune site. The stochastic modeling results showed that most of soil moisture dynamics can be simulated except the rainless periods. Model sensitivity in response to different soil and vegetation parameters was investigated under diverse soil textures. Sensitivity analyses suggested that soil hygroscopic point (sh), field capacity (sfc) were two main parameters controlling the model output. Despite soil moisture dynamics can be partially explained by rainfall, soil moisture dynamics during rainless period still poorly understood. In addition,

characterization of fog distribution in the Namib Desert is still lacking. To this end, nearly two years’ continuous daily records of fog were used to derive fog distribution. The results suggested that fog is able to be well - characterized by a Poisson process with two parameters (arrival rate and average depth). Field observations indicated that there is a moderate positive relationship between soil moisture and fog at GPG and the relationship tend to be less significant at the other two sites. A modified modeling results suggested that mean and general patterns of soil moisture can be captured by the modeling. This thesis is of practical importance for understanding soil moisture dynamics in response to the rainfall and fog changing conditions.

Lixin Wang, Ph. D., Chair

TABLE OF CONTENTS

LIST OF TABLES ...... viii

LIST OF FIGURES ...... ix

Chapter 1: Introduction ...... 1

References ...... 5 Chapter 2: The impact of rainfall on soil moisture dynamics in a foggy desert ...... 9

2.1 Introduction ...... 9 2.2 Materials and Methods ...... 12 2.2.1 Field sites ...... 12 2.2.2 Data collection ...... 13 2.2.3 Data analyses of the field data ...... 14 2.2.4 Model structure and parameterization ...... 15 2.2.5 Model sensitivity analyses ...... 16 2.3 Results and Discussion ...... 17 2.3.1 Field observations ...... 17 2.3.2 Sensitivity analyses ...... 25 2.3.3 Stochastic modeling of soil moisture dynamics ...... 27 2.4 Summary ...... 31 References ...... 32 Chapter 3: The impact of non – rainfall on soil moisture dynamics in the Namib

Desert ...... 39

3.1 Introduction ...... 39 3.2 Material and Methods...... 41 3.2.1 Site descriptions ...... 41 3.2.2 Data collection ...... 43 3.2.3 Analyses of field data ...... 44 3.2.4 Model structure...... 45 3.3 Results and Discussion ...... 46

3.3.1 Fog distribution ...... 46 3.3.2 Field observations ...... 48 3.3.3 Soil moisture modeling with fog as the sole water input ...... 51 3.4 Conclusions ...... 53 References ...... 55 Chapter 4: Conclusions ...... 62

Chapter 5: Appendices ...... 64

Appendix A: Soil moisture dynamics and rainfall patterns at Gobabeb...... 64 Appendix B: Soil moisture dynamics and rainfall patterns at Kleinberg...... 79 Appendix C Soil moisture dynamics and fog patterns at Gobabeb and Kleinberg...... 87 Curriculum Vitae

vii

LIST OF TABLES Table 2.1 Soil moisture and rainfall parameters for different soil depths of Gravel plain (Gobabeb), Sand dune (Gobabeb) and Gravel plain (Kleinberg) Note: The bold letters refer to two sensors at different depths of the same location...... 18

Table 2.2 The correlations between rainfall and soil moisture for different layers of gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 23

Table 2.3 Soil, vegetation and rainfall parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 24

Table 2.4 Model sensitivity of the bounded key parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 25

Table 2.5 Model sensitivity of the non-bounded key parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK). ... 27

Table 2.6 The observed and simulated relative soil moisture (mean ± standard deviation) for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 31

Table 3.1 Soil and fog parameters for gravel plain (Gobabeb)...... 46

viii

LIST OF FIGURES Figure 2.1 Location of the Namib Desert and the Namib “Sand Sea”. Blue points show locations of sites in Gobabeb and Kleinberg. The map was generated using ArcGIS for Desktop 10.3.1 (http://www.arcgis.com)...... 12

Figure 2.2 Rainfall regimes and volumetric soil moisture patterns for different depth of soil types in gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 19

Figure 2.3 Relative soil moisture probability density functions (pdfs) for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 20

Figure 2.4 The comparison between field observations and simulated relative soil moisture patterns in gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK)...... 28

Figure 3.1 Geographic location of study sites (sand dune site at Gobabeb (a), gravel plain site at Gobabeb (b), gravel plain site at Kleinberg (c)) and a schematic photo of fog collector (d). The map was generated using ArcGIS for Desktop 10. 3. 1 1 (http://www.arcgis.com)...... 43

Figure 3.2 Comparison of histograms between field observed fog and Poisson simulated fog at Gobabeb (a) and Kleinberg (b)...... 49

Figure 3.3 Fog events and soil moisture dynamics at sand dune (Gobabeb), gravel plain (Kleinberg) and gravel plain (Gobabeb)...... 50

Figure 3.4 Comparison between field soil moisture dynamics and simulated soil moisture dynamics at gravel plain (Gobabeb, GPG) at the depth of 4 cm...... 52

Chapter 1: Introduction

Defined by their scarcity of precipitation and long dry seasons, drylands are playing an important role in regulating global climate [1]. Drylands cover approximately 40% of the world’s land area and account for 40% of global net primary productivity [2]. However, drylands are facing a number of critical issues (e.g., population growth and water demands) due to anthropogenic influence and the global climate change[3]. Therefore, understanding dryland ecosystems and how non-living components (e.g., rainfall, soil moisture and fog) interact with each other is becoming an international exigency.

Most of the drylands over the world are characterized by their low annual precipitation and high temperature [4]. In this case, dryland soil moisture will be the primary factor that affects and limits development of dryland ecosystem. It has long been suggested that soil moisture is a critical component of dryland ecosystems [1]. Although the amount of dryland soil moisture is relatively small when compared with other humid and semi-humid regions [5, 6], it is of great importance to hydrological, biological and biogeochemical processes [7]. Soil moisture is also a key variable in controlling the exchange of heat fluxes between the land surface and the planetary boundary layer by means of evapotranspiration

[8, 9]. Spatial and temporal variability of dryland soil moisture provides an essential indicator for evaluating and understanding vegetation patterns and dynamics [10]. Soil moisture also controls microbial dynamics and affects a number of soil chemical/physical properties [7], such as O2 levels and pH which in turn affect activities and population dynamics of dryland microbial biomass.

The Namib Desert, which is one of the oldest and largest deserts, is located between a highland plateau and the Atlantic Ocean [11]. It has a total length of 1900 km along the

coast of the Atlantic Ocean, from the Olifants River in South Africa to Carunjamba River in Angola. The annual average rainfall of the Namib Desert is typically low and the distribution is very heterogeneous. The western Namib Desert, on average receives about

5 mm annual rainfall while the eastern part receives about 85 mm [12, 13]. Three main land forms are found in the Namib Desert. The southern Namib Desert is mainly covered by the endless sand dunes called the Namib “Sand Sea”, whereas gravel plains are dominant in the Central Namib dotted with inselbergs of granite and limestone. Moving north from the central Namib Desert, gravels plains finally give way to rugged mountains and dune fields [14]. The hyper-arid environment of the Namib Desert was formed by the cold subantarctic upwelling combined with a hot subtropical interior, resulting in a bleak coastal conditions [15].

In addition to the hyper-arid environment and extremely rare rainfall, the frequent occurrence of fog is the most distinctive characteristic in the Namib Desert. The fog in the

Namib Desert is often considered as a westerly advection fog mainly driven by the Bengula cold current. It has been observed that the Namib Desert fog forms from the coastal areas between midnight and morning and dissipates towards noon. After fog forms at the coastal areas, it is pushed inland by the westerly wind resulting in a west-east gradient foggy zone

[16]. Besides from the Namib Desert, fog also has been recognized as a key water source in other non-dryland ecosystems and its influence on ecological and hydrological processes has brought to a number of scientists’ attention [17-19]. The attention is understandable for researchers who are studying coastal ecology where high frequency of fog was observed during the summer when rainfall is rare and temperature and solar irradiance are highest

[20]. Under this circumstance, many studying has been conducted concerning the effect of

fog on ecological functions. For example, tree physiological responses was found not vary spatially in the rain season, but in the fog season, at the windward edge of the forest versus the interior in the California redwood forest [21]. Moreover, other studies suggested that

Torrey pines on Santa Rosa Island benefit from fog water on a variety of time scales and its growth have a significant positive relationship with the summer fog [22]. In addition to be taken up by trees, researchers also found that fog drip have a significant contribution to groundwater recharge in particular regions. In the study conducted in the Madeira Island, the author assumed that there is a significant difference between isotopic composition in rain and fog (fog being enriched in heavier isotopes 2H and 18O relative to rain) in Madeira

Island. Their results suggested that the isotopic composition of groundwater was in intermediate position between the stable isotopic ratios for rain and fog waters. On top of that, the most possible explanation is that fog water one of water resources that recharges the groundwater system [23].

Despite the recognition of the importance of soil moisture in controlling various ecohydrological processes in the Namib Desert, soil moisture dynamics are poorly reported and how much soil moisture variability can be explained by rainfall and fog in the Namib

Desert have not been reported. This thesis aimed to address the following issues: 1) Fill in a soil moisture data gap in drylands and report a long- term soil moisture dynamics in the

Namib Desert. 2) Use a process-based modeling to simulate soil moisture dynamics under diverse soil textures. 3) Quantify the sensitivity of the stochastic modeling with a range of soil and vegetation parameters. 4) Describe fog statistical distribution in the Namib Desert.

5) Explore the relationships between fog and soil moisture observations in the field.

6) Modify the process-based modeling framework by coupling with non-rainfall component to simulate soil moisture variations during rainfall-free period in the Namib

Desert.

References

[1] Eltahir EAB. A soil moisture rainfall feedback mechanism 1. Theory and observations.

Water Resources Research. 34 (1998) 765-76, doi: 10.1029/97wr03499.

[2] Newman BD, BP Wilcox, SR Archer, DD Breshears, CN Dahm, CJ Duffy, et al.

Ecohydrology of water-limited environments: A scientific vision. Water Resources

Research. 42 (2006), doi: 10.1029/2005wr004141.

[3] Lu X, L Wang, MF McCabe. Elevated CO(2) as a driver of global dryland greening.

Sci Rep. 6 (2016) 20716, doi: 10.1038/srep20716.

[4] Hulme M. Recent climatic change in the world's drylands. Geophysical Research

Letters. 23 (1996) 61-4, doi: 10.1029/95gl03586.

[5] Hollinger SE, SA Isard. A Soil-Moisture Climatology of Illinois. J Climate. 7 (1994)

822-33, doi: 10.1175/1520-0442(1994)007<0822:Asmcoi>2.0.Co;2.

[6] Robock A, KY Vinnikov, G Srinivasan, JK Entin, SE Hollinger, NA Speranskaya, et al.

The Global Soil Moisture Data Bank. B Am Meteorol Soc. 81 (2000) 1281-99, doi:

10.1175/1520-0477(2000)081<1281:Tgsmdb>2.3.Co;2.

[7] Wang LX, S Manzoni, S Ravi, D Riveros-Iregui, K Caylor. Dynamic interactions of ecohydrological and biogeochemical processes in water-limited systems. Ecosphere. 6

(2015) art133, doi:10.1890/Es15-00122.1.

[8] Seneviratne SI, T Corti, EL Davin, M Hirschi, EB Jaeger, I Lehner, et al. Investigating soil moisture-climate interactions in a changing climate: A review. Earth-Science Reviews.

99 (2010) 125-61, doi: 10.1016/j.earscirev.2010.02.004.

[9] D'Odorico P, L Ridolfi, A Porporato, I Rodriguez-Iturbe. Preferential states of seasonal soil moisture: The impact of climate fluctuations. Water Resources Research. 36 (2000)

2209-19, doi: 10.1029/2000wr900103.

[10] Ravi S, P D'Odorico, L Wang, S Collins. Form and function of grass ring patterns in arid grasslands: the role of abiotic controls. Oecologia. 158 (2008) 545-55, doi:

10.1007/s00442-008-1164-1.

[11] Stone AEC. Age and dynamics of the Namib Sand Sea: A review of chronological evidence and possible landscape development models. Journal of African Earth Sciences.

82 (2013) 70-87, doi: 10.1016/j.jafrearsci.2013.02.003.

[12] Eckardt FD, K Soderberg, LJ Coop, AA Muller, KJ Vickery, RD Grandin, et al. The nature of moisture at Gobabeb, in the central Namib Desert. Journal of Arid Environments.

93 (2013) 7-19, doi: 10.1016/j.jaridenv.2012.01.011.

[13] Henschel JR, MK Seely. Ecophysiology of atmospheric moisture in the Namib Desert.

Atmospheric Research. 87 (2008) 362-8, doi: 10.1016/i.atmosres.2007.11.015.

[14] Bristow CS, GAT Duller, N Lancaster. Age and dynamics of linear dunes in the Namib

Desert. Geology. 35 (2007) 555-8, doi: 10.1130/G23369a.1.

[15] Hachfeld B, N Jurgens. Climate patterns and their impact on the vegetation in a fog driven desert: The Central Namib Desert in Namibia. Phytocoenologia. 30 (2000) 567-89, doi: 10.1127/phyto/30/2000/567.

[16] Kaseke KF, AJ Mills, K Esler, J Henschel, MK Seely, R Brown. Spatial Variation of

“Non-Rainfall” Water Input and the Effect of Mechanical Soil Crusts on Input and

Evaporation. Pure and Applied Geophysics. 169 (2012) 2217-29, doi: 10.1007/s00024-

012-0469-5.

[17] Burgess SSO, TE Dawson. The contribution of fog to the water relations of Sequoia sempervirens (D. Don): foliar uptake and prevention of dehydration. Plant Cell Environ.

27 (2004) 1023-34, doi: 10.1111/j.1365-3040.2004.01207.x.

[18] Dawson TE. Fog in the California redwood forest: ecosystem inputs and use by plants.

Oecologia. 117 (1998) 476-85, doi: 10.1007/s004420050683.

[19] Williams AP, RE Schwartz, S Iacobellis, R Seager, BI Cook, CJ Still, et al.

Urbanization causes increased cloud base height and decreased fog in coastal Southern

California. Geophysical Research Letters. 42 (2015) 1527-36, doi: 10.1002/2015gl063266.

[20] Ingraham NL, RA Matthews. The Importance of Fog-Drip Water to Vegetation - Point-

Reyes Peninsula, California. J Hydrol. 164 (1995) 269-85, doi: 10.1016/0022-

1694(94)02538-M.

[21] Ewing HA, KC Weathers, PH Templer, TE Dawson, MK Firestone, AM Elliott, et al.

Fog Water and Ecosystem Function: Heterogeneity in a California Redwood Forest.

Ecosystems. 12 (2009) 417-33, doi: 10.1007/s10021-009-9232-x.

[22] Williams AP, CJ Still, DT Fischer, SW Leavitt. The influence of summertime fog and overcast clouds on the growth of a coastal Californian pine: a tree-ring study. Oecologia.

156 (2008) 601-11, doi: 10.1007/s00442-008-1025-y.

[23] Susana Prada, José V. Cruz, Manuel O. Silva, Celso Figueira. Contribution of cloud water to the groundwater recharge in Madeira Island: preliminary isotopic data.

Chapter 2: The impact of rainfall on soil moisture dynamics in a foggy desert

2.1 Introduction

It has long been suggested that soil moisture is a critical component of earth systems

[1]. Although the amount of soil moisture is relatively small when compared with other constituents of the hydrological cycle [2, 3], it is of great importance to many hydrological, biological and biogeochemical processes. Soil moisture, as one source of water for the atmosphere through evapotranspiration [4], is a key variable in controlling the exchange of heat fluxes between the land surface and the planetary boundary layer [5]. Spatial and temporal variability of soil moisture provides an essential indicator for evaluating and understanding vegetation patterns and dynamics [6, 7]. Soil moisture also controls microbial dynamics and affects a number of soil chemical/physical properties [8], such as

O2 levels, pH and the concentration of mineral nutrients (e.g., ferric iron) in soil solution, which in turn affect activities and population dynamics of microbial biomass.

Soil moisture is especially important to link climate, soil, and vegetation in dryland ecosystems. Drylands cover about 40% of the earth surface and are typically located in continental regions where rainfall is less than potential evapotranspiration [9-11]. Dryland soil water content is typically low, but soil water is critical for vegetation dynamics and moisture in the topsoil can effectively protect the dryland soil from wind erosion [12, 13].

Typically, precipitation is the major source of soil moisture in drylands, though in some fog dominated systems, the non-rainfall water (e.g., fog and dew) can exceed annual rainfall [14]. Dryland near-surface climate is affected by soil moisture, which has been revealed as a major factor contributing to the occurrence of extremely high temperature and drought weather [10]. Therefore, understanding interactions between soil moisture and

precipitation is critical to predict the response of dryland ecosystems to global environmental changes. To make a long-term prediction of soil moisture, process-based modeling is often required. One of the recent advances in ecohydrology is the successful modeling of stochastic rainfall-soil moisture relationships. Recently a modeling framework was developed based on the stochastic characteristic of rainfall events and analytical results of the probability distributions of soil moisture were successfully obtained. The modeling framework was improved and modified in a previous study to achieve a more accurate description of soil moisture dynamics especially in water-limited systems [15]. A series of studies have applied the modeling framework to different dryland environments [16-18].

Although the modeling framework has been used for years, the sensitivities of the key parameters and their dependence on soil texture are poorly understood. The hyper-arid environment of the Namib Desert and the diverse soil textures under similar rainfall regimes make it an ideal place to test the sensitivities across various soil textures.

The Namib Desert, which is one of the oldest and largest deserts, is located between a highland plateau and the Atlantic Ocean [19]. The hyper-arid environment of the Namib

Desert was formed by the cold subantarctic upwelling combined with a hot subtropical interior, resulting in a bleak coastal condition [20, 21]. The annual average rainfall of the

Namib Desert is typically low and the distribution is very heterogeneous. The western

Namib Desert, on average receives about 5 mm annual rainfall while the eastern part receives about 85 mm [22, 23]. In addition to the hyper-arid environment and extremely rare rainfall, the frequent occurrence of fog is the most distinctive characteristic in the

Namib Desert [21]. The fog in the Namib Desert is often considered as a westerly advection fog mainly driven by the Bengula cold current [24]. It has long been observed that the

Namib Desert fog forms from the coastal areas between midnight and morning and dissipates towards noon. After fog forms at the coastal areas, it is pushed inland by the westerly wind resulting in a west-east gradient foggy zone [25]. The Namib Desert fog as a source of water has been playing an important role in sustaining plant growth by means of interception and can also be used for the survival of small animals [26, 27]. For example, fog water uptake has been observed for three beetle species of the Namib Desert when the beetles face extreme surface temperature and wind [28]. The Namib grass Stipagrostis sabulicola was found to rely heavily on fog water to sustain themselves and is able to transfer fog water intercept by leaves to their plant base by stem flow [29]. It has also been reported that 19% of the water within the Sequoia sempervirens, and 66% of the water within the understory plants come from fog in the California redwood forests [30, 31].

Despite the recognition of the importance of soil moisture in controlling various ecohydrological processes in the Namib Desert, soil moisture dynamics and how much soil moisture variability can be explained by rainfall in the foggy Namib Desert have not been reported. In this study, twelve to nineteen months’ daily records of rainfall and soil moisture records from diverse ecosystems in the Namib Desert were reported. The objects of this study are to, 1) present field observations of soil moisture and rainfall records acquired from different soil types in the Namib Desert; 2) use process-based modeling to simulate soil moisture dynamics under diverse soil textures; and 3) quantify the sensitivity of the stochastic modeling with a range of soil and vegetation parameters.

2.2 Materials and Methods

2.2.1 Field sites

The Namib Desert is an ancient desert located in the coastal area of Namibia. It has a total length of 1900 km along the coast of the Atlantic Ocean, from the Olifants River in

South Africa to Carunjamba River in Angola (Fig. 2.1). The average rainfall in the Namib

Desert ranges from 50 to 100 mm in the far south, 5-18 mm in the central Namib Desert and less than 50 mm along the Angolan coast in the north [22, 26]. Three main land forms are found in the Namib Desert. The southern Namib Desert is mainly covered by the endless sand dunes called the Namib “Sand Sea”, whereas gravel plains are dominant in the Central Namib dotted with inselbergs of granite and limestone. Moving north from the central Namib Desert, gravels plains finally give way to rugged mountains and dune fields

[32].

Figure 2.1 Location of the Namib Desert and the Namib “Sand Sea”. Blue points show locations of sites in Gobabeb and Kleinberg. The map was generated using ArcGIS for Desktop 10.3.1 (http://www.arcgis.com).

Gobabeb is located at the edge of the Namib “Sand Sea” (Fig. 2.1), 60 km inland from the Atlantic Ocean [22]. The climate is hyper-arid with average annual rainfall less than 50 mm primarily concentrated around January to March. Two different landscapes occur near

Gobabeb along both sides of the Kuiseb River, the “Sand Sea” to the south and the gravel plains (calcrete soil) to the north [33]. Major plant species in the gravel plain are

Zygophyllum simplex and Z. stapffi, while Stipagrostis sabulicola and Trianthema heroensis are the most commonly seen species dotted in the sand dune area [34]. Kleinberg

(Fig. 2.1), located about 33 km from the Atlantic Ocean, has been a field site attached to the Gobabeb Research and Training Centre for over 30 years, with most of the area covered by gravel plains (gypsum soil) [33] dotted with pencil bush (Arthraerua leubnitziae) and assorted lichens. The total annual rainfall amounts in both Gobabeb and Kleinberg are similar with rainfall events rarely occur.

2.2.2 Data collection

Three sites (Kleinberg gravel plain, here after GPK; Gobabeb sand dune (High Dune), here after SDG; Gobabeb gravel plain, here after GPG) were selected in our study because of the similar meteorological conditions and different soil textures among these three sites.

Twelve to nineteen months’ volumetric soil moisture data and the corresponding rainfall data (January 1, 2014 to August 3, 2015 for GPK; July 28, 2014 to July 28, 2015 for SDG;

January 2, 2014 to July 28, 2015 for GPG) were collected and used to test the stochastic modeling. The data collection was granted with a research permit from Gobabeb Research and Training Center of Namibia. Daily rainfall data were obtained from two tipping- buckets (one at GPG and another at GPK), which have been calibrated in the field. The

same tipping bucket data was used for GPG and SDG because of their proximity (about

3.5 km apart). Soil moisture from both bare soil and vegetated areas were continuously monitored at one-hour interval using CS655 Water Content Reflectometer (Campbell

Scientific, Inc. Logan, Utah, USA). In total six probes were used to measure soil moisture under different layers at the three sites. Three probes were used at GPG with two being installed at the same location under bare soil at 7.5 cm and 22.5 cm and the other one being installed under vegetation cover at 7.5 cm. Two probes were installed at SDG with one under bare soil at 15 cm and another one under vegetation cover at 7.5 cm. The depths of the two probes at the sand dune site were different due to the movement of shifting sand.

One probe was installed at GPK under bare soil at 5 cm. Saturated hydraulic conductivity

(Ks) were estimated using the mini disk tension infiltrometer (Decagon Inc. Pullman, WA,

USA) from multiple locations at each site.

2.2.3 Data analyses of the field data

Hourly volumetric soil moisture data were averaged to daily scale in order to match the model time scale. Data central tendency and variability were reported as mean, standard deviation, and coefficient of variation. The Mann–Whitney U test was deployed to examine the differences of mean soil moisture among the three sites in PAST (Paleontological

Statistics, Natural History Museum, University of Oslo). Surface soil moisture distribution was examined using Q-Q/P-P plot in IBM SPSS (IBM Inc. NY, USA) and the corresponding probability density functions (pdfs) were described as Gamma distributions using two parameters (shape parameter k and scale parameter θ) in MATLAB. Correlations between rainfall events and soil moisture were tested using Pearson’s correlation.

2.2.4 Model structure and parameterization

Field soil moisture observations were modeled at daily scale by utilizing a process- based stochastic model. The model was defined with rainfall inputs as a non-homogeneous

Poisson process with storm arrival rate and mean event depth α. However, in this study, we used a deterministic approach using field rainfall data to drive the model. Soil moisture dynamics are expressed as following equations:

푑푠 푛푍 = 퐼(푡, 푠) − 퐸푇(푠) − 퐿(푠), (1) 푟 푑푡 where n is the porosity, s is the relative soil moisture, Zr is active soil depth or rooting depth,

I(s, t) is the infiltration rate of rainfall, ET(s) is the rate of evapotranspiration, and L(s) is the rate of leakage or the loss of soil moisture from the bottom layer. The model assumes that all the water input from rainfall is immediately infiltrated into the ground and no surface runoff is generated.

There are three major factors in equation (1) controlling soil moisture content. The increase of soil moisture is due to infiltration and the decrease of soil moisture is due to evapotranspiration and leakage through the bottom layer. The combined effect of those processes can be described as:

푠∗−s 퐸 + (퐸 − 퐸 ) 푠 < 푠 ≤ 푠∗ vap max vap 푠∗−푠 w w 퐿(푠) + 퐸푇(푠) = ∗ , (2) 퐸max 푠 < 푠 ≤ 푠fc 퐾 s 훽(1−푠fc) 퐸max + (e − 1) 푠fc < 푠 ≤ 1 { e훽(1−푠fc) where sw is the soil water content at wilting point, sfc is the soil water content at field capacity, s* is the soil moisture in conditions of incipient stress, Ks is the hydraulic conductivity, β = 2b + 4, with b being the pore size distribution index; Evap is the evaporation rate of ground surface while Emax being the maximum evapotranspiration

under well-watered condition. For s > sfc, losses of soil moisture come from evapotranspiration and leakage. For s* < s ≤ sfc, only evapotranspiration contributes to

the loss of soil water at the maximum evapotranspiration rate Emax. For sw < s ≤ s*, vegetation begins to suffer from water stress and regulates the transpiration rate through stomata closure. Thus, transpiration starts to be limited by soil moisture and the total evapotranspiration also decreases with decreasing soil moisture. For s < sw, ET(s) linearly decreases reaching a zero at s = sh (hygroscopic point) where soil begins to absorb water from the atmosphere.

2.2.5 Model sensitivity analyses

Sensitivity analyses were conducted to examine the response of the modeled soil moisture output to soil and vegetation parameters under different soil textures. Sensitivity analyses were conducted by changing one parameter while fixing others (i.e., no interactive effects were tested). Before sensitivity analyses, we divided the key model parameters into bounded group (e.g., porosity (n), field capacity (sfc)) and non-bounded group (soil depth

(Zr), saturated hydraulic conductivity (Ks)) in terms of whether they have been normalized to the range of 0 to 1 or not. The maximum values of the bounded group were all set to 1, while the maximum values of the non-bounded group were set to different values. We predefined 20 as the maximum value for Zr, Tmax and Ew. Meanwhile, 100 was set to be the maximum value for Ks considering its magnitude in reality. Then in order to ensure the accuracy and precision of our sensitivity analyses, the parameter ranges were further refined to determine new parameter ranges based on the model outputs obtained from the

previous procedures. Finally, based on the curve shape of the model output, we divided them into three groups: (1) monotonic increasing group, (2) monotonic decreasing group, and (3) non-monotonic group. For monotonic groups, the minimum values of the parameter ranges were defined as values that start making average-simulated soil moisture greater than zero while the maximum values were defined as values after which the average- simulated soil moisture will level off within the predefined or normalized ranges. For the non-monotonic group, the largest monotonic range (either monotonic increasing or decreasing) within the predefined or normalized range was regarded as the final range of the parameter. The average of the difference between two consecutive soil moisture output values divided by the parameter increment was then define as the parameter’s sensitivity, which can be described as:

∑푛 (푠(𝑖)−푠(𝑖−1)) 푆푒푛푠푖푡푖푣푖푡푦 = 𝑖=2 ∗ 100% , (3) (푛−1) ∗ 푥 where s (i) is the correspondent soil moisture value, s (i-1) is soil moisture value one increment before s (i) produced within the predefined parameter range, n is the number of values, which equals to the determined parameter range divided by the parameter increment, x is parameter increments (e.g., 0.001 for porosity).

2.3 Results and Discussion

2.3.1 Field observations

Table 1 shows the rainfall parameters and mean soil moisture values measured at the three sites. It was observed that total rainfall at GPG and SDG was 15.75 mm in 2014, which was slightly less than that at GPK (18.45 mm, Table 2.1). Despite average rainfall depth for GPG/SDG (2.63 mm) was slightly higher than that of GPK (1.95 mm), rainfall

frequency at GPG/SDG (0.01) was only one quarter of that in GPK (0.04) (Table 2.1).

Although divergences of rainfall parameters and total annual rainfall amount existed among these sites, the rainfall patterns were generally the same (Fig. 2.2). Most of the rainfall concentrated on the wet season from November to May and large rainfall events mainly occurred in January (Fig. 2.2).

Table 2.1 Soil moisture and rainfall parameters for different soil depths of Gravel plain (Gobabeb), Sand dune (Gobabeb) and Gravel plain (Kleinberg) Note: The bold letters refer to two sensors at different depths of the same location. Bare Mean soil moisture CV Rainfall α Study site λ soil/Vegetation (%) (%) (mm) (mm) Gravel plain Bare soil (7.5cm) 1.97±0.39 19.79 15.75 0.01 2.63 (Gobabeb) Bare soil (22.5cm) 0.62±0.19 30.65 15.75 0.01 2.63 Vegetation (7.5cm) 2.69±1.07 39.78 15.75 0.01 2.63 Sand dune Bare soil (15cm) 0.57±0.12 21.05 15.75 0.01 2.63 (Gobabeb) Vegetation (7.5cm) 0.73±0.08 10.96 15.75 0.01 2.63 Gravel plain Bare soil (5.0cm) 0.70±0.40 57.14 18.45 0.04 1.95 (Kleinberg)

Figure 2.2 Rainfall regimes and volumetric soil moisture patterns for different depth of soil types in gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK).

Figure 2.3 Relative soil moisture probability density functions (pdfs) for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK).

Fig 2.3 shows the simulated soil moisture probability density functions (pdfs) of GPG

(bare soil 7.5 cm), GPK (bare soil 5 cm) and SDG (bare soil 15 cm) based on field measurements. Soil moisture pdf shape of GPG was different from that of SDG, which had smaller mode value and longer tail. The difference can be directly reflected by the values

of shape parameter k and scale parameter θ, with GPG having a higher Ks (43.1) and lower

θ (1.3 x 10-3). The discrepancy of soil moisture pdfs between GPG and SDG may result from the different antecedent soil moisture and the similarity of rainfall patterns (Fig. 2.2).

Soil moisture pdfs of SDG and GPK shared some similarities, with relatively low k (20.0 and 5.9, respectively) and high θ (7.2 and 2.3 x 10-3, respectively). The similarities may be explained by the proximity of their initial soil moisture of SDG and GPK and the resemblance of soil moisture mode value (Fig. 2.2).

Significant differences of soil moisture dynamics can be found at different depths of soil layers at GPG (Fig. 2.2) and soil moisture differences became larger after storms. Soil moisture difference between layers might be explained by the high dependence of surface soil moisture on the prevailing environment since it is continuously gaining and loosing water by means of rainfall infiltration and evapotranspiration. The phenomenon is more pronounced in the unsaturated zone or drylands where soil moisture largely depends on meteorological conditions [35, 36]. This behavior is also consistent with observations from other studies in arid regions [37, 38] but differs from the results obtained in China where researchers found no significant differences for soil water content at 0.4 m, 0.6 m and 0.8 m in a small watershed [39]. Besides, soil moisture differences between different layers were even found to be smaller after storms [40, 41]. The consistency between our results and previous research may be explained by the similarities of meteorological conditions.

In general, low and irregular annual rainfall tend to result in high soil moisture in the surface soil and low soil moisture in the deep soil layer because when rainfall amount is too small, most of them will be retained in the shallow layer. That may help explain why soil moisture differences between surface layer and deeper layer tend to be larger after

storms in our study sites. The discrepancy between our results and other studies may be induced by the difference in soil texture and hydrological conditions. Soils in other study sites may be able to hold less moisture, or have higher infiltrate rate or have more intense interaction with the groundwater, which may result in insignificant differences between shallower and deeper layers event after a strong storm. Soil moisture under vegetation cover was higher than that of bare soil at GPG (p < 0.05) (Fig 2.2, Table 2.1). This phenomenon is also suggested by other dryland studies [42]. This might be explained by the fact that bare soils tend to dry out faster due to the higher solar energy they receive. In addition, the net gain between retaining infiltrating rainfall and soil moisture losses through evapotranspiration under vegetated soil is likely larger than bare soil moisture loss through evaporation in water-limited systems. The combination of those two effects will result in higher soil moisture under vegetated soil and lower soil moisture in bare soils. However, this was not universal, a negative relationship was found between soil moisture and canopy cover at low soil moisture and the negative relationship diminished when soil became wetter [43]. More intense temporal soil moisture fluctuations can be found in the top soil layer than in the deeper layers. This result is in good agreement with the observations in

Wagga Wagga and Tarrawarra [44] and soil moisture fluctuations may became less pronounced in deeper soil layers as suggested by recent studies in China [45, 46]. At GPG, mean soil moisture and standard deviation in the top layer (1.97% and 0.39% respectively) were more than two times that in the deeper layer (0.62% and 0.19% respectively), whereas the CV of the deeper layer (30.65%) was nearly double that in the top layer (19.79%) (Table

2.1). The soil moisture difference between top layer and deeper layer in at GPG (bare soil) might be induced by the soil properties. As suggested by previous studies, soils at the gravel

plains around GPG are primarily calcrete [33, 36]. A study conducted in semiarid southern

New Mexico suggested that caliche can absorb considerable amount of water [47] which is not easy to be released to the surrounding soil or taken up by vegetation.In addition, calcrete soil has low infiltration capacity limiting water movement toward into deeper soil layers. In SDG, there was no significant difference (p < 0.05) in average soil moisture between bare soil (0.57%) and under vegetation cover (0.73%) microsites (Table 2.1). This is consistent with a previous study which found the average soil moisture is nearly the same in the two microsites (bare soil and covered with vegetation, respectively) along the

Kalahari Transect [17]. The soil moisture standard deviations were very similar in both microsites, while their CVs were different with bare soil (20.05%) being two times that of microsites covered with vegetation (10.96%). At GPK, the mean soil moisture (0.70%) was similar to the soil moisture for bare soil at SDG (p < 0.05) (Table 2.1). This can be clearly reflected by the soil moisture pdfs at GPK and SDG from which nearly the same shapes have been observed (Fig 2.3). The CV (57.14%) for GPK was the highest when comparing with the microsites at GPG and SDG.

Table 2.2 The correlations between rainfall and soil moisture for different layers of gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK).

Study site Bare soil/Vegetation Depth (cm) r P

Gravel plain (Gobabeb) Bare soil 7.5 0.49 0 Bare soil 22.5 0.03 0.53 Vegetation 7.5 0.03 0.52 Sand dune (Gobabeb) Bare soil 15 0.07 0.17 Vegetation 7.5 0.01 0.82 Gravel plain (Kleinberg) Bare soil 5 0.39 0

At GPG and GPK, soil moisture always followed the rainfall patterns with soil moisture peaks following rainfall events, whether the surface was covered with vegetation or not (Fig 2.2). At the top soil layer, positive correlations between rainfall and soil moisture dynamics were more obvious (Table 2.2). However, soil moisture dynamics did not have any correlation with rainfall at SDG (Table 2.2). The inconsistencies between the three sites were presumably caused by the infiltration capacity. In our field measurements, higher Ks (50.6 m day-1) (Table 2.3) was observed at SDG, which was nearly one order of magnitude higher than that of GPG (5.6 m day-1) and GPK (3.5 m day-1) (Table 2.3). This indicates that water would leach away at a fast rate at SDG and can hardly be retained by the soil, whereas some water might be captured by soil at GPG and GPK after rainfall events.

Table 2.3 Soil, vegetation and rainfall parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK). Unit Gravel plain Sand dune Gravel plain (Gobabeb) (Gobabeb) (Kleinberg) Bare soil Bare soil Bare soil

Soil parameters Porosity n 0.34 0.40 0.48 Field capacity sfc 0.10 0.05 0.09 Hygroscopic point sh 0.05 0.015 0.01 Saturated hydraulic -1 conductivity Ks (m day ) 5.60 50.60 3.50 Soil depth Zr (m) 0.21 0.48 0.35 Rainfall parameters -1 Average storm frequency λ (day ) 0.01 0.01 0.04 Average storm depth α (mm) 2.62 3.02 1.95 Vegetation parameters -1 Maximum evapotranspiration Emax (mm day ) 1.25 1.15 2.20 Soil-vegetation parameters Point of incipient stress s* 0.09 0.03 0.08 Permanent wilting point sw 0.075 0.020 0.045

2.3.2 Sensitivity analyses

The modeled soil moisture sensitivities to the model parameters from both bounded and non-bounded groups were tested. In general, the model was more sensitive to the parameters in the bounded group with the average sensitivity ranging from 0.00011% to

44%. In comparison, the average sensitivity ranged from -0.00065% to 0.07% for the non- bounded group (Table 2.4). For parameters in the bounded group, the sensitivities were all positive except n at SDG. This indicates that simulated soil moisture will increase as values of these parameters become higher. In the non-bounded group, however, nearly all the parameters had negative sensitivity except Zr at GPG and GPK.

Table 2.4 Model sensitivity of the bounded key parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK). Porosity Field Hygroscopic Point of Permanent incipient wilting capacity point stress point

Unit n Sfc Sh Ks Sw

Step 0.001 0.001 0.001 0.001 0.001

0.002, 0.001, 0.001, 0.001, 0.001, Interval GPG 0.981 0.156 0.094 0.985 1 Average 0.75 15 44 0.75 0.086 sensitivity (%)

Step 0.001 0.001 0.001 0.001 0.001

0.006, 0.001, 0.001, 0.001, 0.001, Interval GPK 0.969 0.585 0.404 0.847 1 Average 0.00011 1.4 11 0.17 2.6 sensitivity (%)

Step 0.001 0.001 0.001 0.001 0.001

0.082, 0.001, 0.001, 0.001, 0.001, Interval SDG 0.899 0.04 0.059 0.984 0.964 Average -0.31 1.9 2.1 0.37 3.2 sensitivity (%)

Sensitivity analyses suggested that parameter sensitivities depended on the overall model parameterization, thus the parameter sensitivity was different for each site. For the bounded group at GPG, average sensitivities of sh, sfc, n and s* were of the same order of magnitude. Among the bounded group, sh had the largest average sensitivity thousands of times more than the average sensitivity of sw. In the non-bounded group, the vegetation parameter Tmax and Ew exhibited the same sensitivity value of -0.11% while average sensitivity values of soil parameter Zr and Ks were quite different, with 0.07% for Zr and -

0.00065% for Ks (Table 2.5). For all parameters at GPK, sh was the most influential factor for the model among all the parameters with an average sensitivity value of 15%. In the bounded group, the minimum average sensitivity was n, which was thousands of times smaller than that at GPG. In the non-bounded group, all the average sensitivity values were negative except Zr, which was far larger than any other values of the group. With respect to SDG, sh had the largest average sensitivity value and it had the same magnitude with sfc, which was very similar to that at GPG (Table 2.4). But the average sensitivity of n had a totally distinct trend to that of the other two sites suggesting that simulated soil moisture will decrease as n goes higher within the predefined interval. The same patterns happened in the non-bounded group in which the average sensitivity of Zr was -0.036% (Table 2.5), though the absolute value of Zr had the same magnitude as those at the other two sites. For average sensitivities of the three sites, the model was more sensitive to sh and sfc and less sensitive to Ks. All the soil parameters had positive values except n, Zr of SDG and Ks of all the sites whereas all the vegetation parameters had negative values.

Table 2.5 Model sensitivity of the non-bounded key parameters for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK). Soil Saturated Maximum Maximum depth hydraulic transpiration evaporation conductivity

n Tmax Ew Ks Unit m mm day-1 mm day-1 m day-1

Step 0.01 0.01 0.01 0.1

0.01, 0.01, 0.001, 0.1, Interval 15.5 11.04 13.75 81.4 GPG Average sensitivity 0.07 -0.11 -0.11 -0.00065 (%)

Step 0.01 0.01 0.01 0.1

0.06, 0.01, 0.01, 0.1, Interval 15.83 11.13 8.36 81.4 GPK Average sensitivity 0.032 -0.036 -0.0014 -0.0044 (%)

Step 0.01 0.01 0.01 0.1

0.01, 0.01, 0.01, 0.1, Interval 18.11 7.77 4.27 100 SDG Average sensitivity -0.036 -0.06 -0.057 -0.00041 (%)

2.3.3 Stochastic modeling of soil moisture dynamics

Soil moisture from the three sites (7.5 cm bare soil at GPG; 15 cm bare soil at SDG and 5 cm at GPK) were selected and simulated by a stochastic modeling framework. The modeled mean relative soil moisture and soil moisture dynamics in general agreed well with field observations (Fig 2.4).

Figure 2.4 The comparison between field observations and simulated relative soil moisture patterns in gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK).

Simulated mean relative soil moisture at the three sites was 0.056 for GPG, 0.014 for

SDG and 0.015 for GPK (Table 2.6) and mean soil moisture observed in the field was 0.058,

0.014 and 0.015 for GPG, SDG and GPK, respectively (Table 2.6). Soil moisture at GPG can be well simulated, with soil moisture peaks corresponding to rainfall regimes. Fig. 2.4

shows that the simulated GPG soil moisture was slightly lower than soil moisture acquired from the field, with soil moisture increasing sharply to the peak during a rainfall event, and decreasing rapidly to a soil moisture baseline right after a rainfall event. The sharp decrease of soil moisture might be explained by the modeling assumption that rainfall is the only contributor to the increase of soil moisture. When intense rainfall inputs were added, soil moisture would immediately increase to field capacity and leakage would instantaneously occur until soil moisture returned below field capacity. For SDG, intense soil moisture peaks can be observed from simulated results, with each soil moisture peak directly following the rainfall patterns (Fig 2.4). Although rainfall patterns at SDG are the same as that in GPG, the soil moisture simulation results at SDG were quite different. Soil moisture of SDG suddenly increased to a fixed value when a storm came and stayed at that value until another rainfall came which is different from the persistent decrease of soil moisture at GPG. This is because our initial soil moisture at SDG was below sh and the first rainfall may not have been sufficient for soil to reach its hygroscopic point. This caused a flat and a sharp increase of soil moisture in the beginning. On the other hand, simulated soil moisture of SDG always reached the baseline soil moisture without any smooth soil moisture transition or soil moisture fluctuation. In contrast, the soil moisture curve of GPG reached the baseline value very smoothly, though still without any soil moisture fluctuation

(Fig. 2.4). The differences may be induced by soil properties where larger Ks and smaller sfc was found in SDG (Table 2.3). For GPK, fluctuations can be seen when rainfall came.

At the beginning, simulated soil moisture was slightly higher than the measured soil moisture value but after two intense rainfall events the simulated results were smaller than field observations (Fig. 2.4). In general, soil moisture patterns and mean relative soil

moisture can be well-simulated by the stochastic model. However, daily soil moisture fluctuations cannot be fully revealed by the model simulation. From our perspective, two factors mainly influence the model output. The first factor is the effect of non-rainfall components, particularly fog and dew, which influence the daily soil moisture fluctuations while the model failed to take them into account resulting in steep slopes of soil moisture of two adjacent days between and after a rainfall event. It has been suggested that fog has been persistent in Namib Desert [48]. High fog is the most common type of fog that can be found at Gobabeb, which arrives during early morning hours and dissipates quickly with sunrise when the surface temperature rises. This is probably the first reason why we cannot simulate the daily soil moisture fluctuations. In addition to the non-rainfall component effect, the modeling framework itself is based on the assumption that daily scale infiltration and redistribution occur instantaneously, and soil moisture that exceeds field capacity will be drained away immediately [49]. Moreover, the model does not consider vertical distribution of soil moisture, assuming soil moisture is the same along soil columns. In reality, however, soil moisture is different from one layer to another, with soil moisture in a shallow layer generally having higher soil water content than that in a deep layer.

Our simulation results showed that the stochastic model can be used to simulate soil moisture patterns in the Namib Desert especially in gravel plains where finer soil texture was found. In the site with coarse soil texture (e.g., SDG), the model did not perform very well, indicating that the modeling framework may not be able to accurately predict soil moisture dynamics at daily scale for sites with coarse texture.

Table 2.6 The observed and simulated relative soil moisture (mean ± standard deviation) for gravel plain at Gobabeb (GPG), sand dune at Gobabeb (SDG) and gravel plain at Kleinberg (GPK). Gravel plain Sand dune Gravel plain Study sites Gobabeb Gobabeb Kleinberg

Soil type Bare soil Bare soil Bare soil

Depth (cm) 7.5 15 5.0

Observed 0.058±0.010 0.014±0.003 0.015±0.008

Simulated 0.056±0.011 0.014±0.006 0.015±0.008

2.4 Summary

In this study, twelve to nineteen months’ daily-scale soil moisture and rainfall data were obtained from three sites located within the Namib Desert. The ground observations showed that soil moisture was controlled by rainfall patterns at GPG and GPK, particularly for shallow soil layers with strong correlations between soil moisture and rainfall, while weak rainfall-soil moisture correlation was found at the sand dune site. The field observations were simulated using a process-based modeling framework. The modeled soil moisture patterns and mean soil moisture values agreed well with field observations.

However, soil moisture fluctuations cannot be simulated and require future work such as taking fog and dew into consideration as additional water inputs. The model sensitivity showed that sensitivity patterns were quite similar between the three sites. But the sensitivity magnitude of the model parameters differed from each other, with sh and sfc having the largest sensitivities among all the parameters. The sensitivity analyses of the three sites were quantified and can be used as an uncertainty indicator for this modeling framework in future applications.

References

[1] Eltahir EAB. A soil moisture rainfall feedback mechanism 1. Theory and observations.

Water Resour Res. 1998;34(4):765-76. doi: 10.1029/97wr03499

[2] Hollinger SE, Isard SA. A soil-moisture climatology of Illinois. J Climate.

1994;7(5):822-33. doi: 10.1175/1520-0442(1994)007<0822:ASMCOI>2.0.CO;2

[3] Robock A, Vinnikov KY, Srinivasan G, Entin JK, Hollinger SE, Speranskaya NA, et al. The global soil moisture data bank. B Am Meteorol Soc. 2000;81(6):1281-99. doi:10.1175/1520-0477(2000)081<1281:Tgsmdb>2.3.Co;2

[4] Seneviratne SI, Corti T, Davin EL, Hirschi M, Jaeger EB, Lehner I, et al. Investigating soil moisture-climate interactions in a changing climate: A review. Earth-Sci Rev.

2010;99(3-4):125-61. doi: 10.1016/j.earscirev.2010.02.004

[5] D'Odorico P, Ridolfi L, Porporato A, Rodriguez-Iturbe I. Preferential states of seasonal soil moisture: the impact of climate fluctuations. Water Resour Res. 2000;36(8):2209-

19.doi: 10.1029/2000WR900103

[6] Ravi S, D'Odorico P, Wang L, Collins S. Form and function of grass ring patterns in arid grasslands: the role of abiotic controls. Oecologia. 2008;158(3):545-55. doi:

10.1007/s00442-008-1164-1

[7] Yu KL, D'Odorico p. An ecohydrological framework for grass displacement by woody plants on savannas. J Geophys Res-Biogeo. 2014:119:192-206. doi:

10.1002/2013JG002577

[8] Wang LX, Manzoni S, Ravi S, Riveros-Iregui D, Caylor K. Dynamic interactions of ecohydrological and biogeochemical processes in water-limited systems. Ecosphere.

2015;6(8):1-27. doil:10.1890/ES15-00122.1

[9] Wang LX, D'Odorico P, Evans JP, Eldridge DJ, McCabe MF, Caylor KK, et al. Dryland ecohydrology and climate change: critical issues and technical advances. Hydrol Earth Syst

Sc. 2012;16(8):2585-603. doi:10.5194/hess-16-2585-2012

[10] D'Odorico P, Porporato A. Dryland ecohydrology. Dordrecht, Netherlands: Springer;

2006.

[11] D'Odorico P, Bhattachan A, Davis KF, Ravi S, Runyan CW. Global desertification:

Drivers and feedbacks. Adv Water Resour. 2013; 51:326-44. doi:

10.1016/j.advwatres.2012.01.013

[12] Ravi S, D'Odorico P, Wang LX, White CS, Okin GS, Macko SA, et al. Post-fire resource redistribution in desert grasslands: a possible negative feedback on land degradation. Ecosystems. 2009; 12(3):434-44. doi: 10. 1007/s10021-009-9233-9

[13] Wang LX, D'Odorico P, O'Halloran LR, Caylor K, Macko S. Combined effects of soil moisture and nitrogen availability variations on grass productivity in African savannas.

Plant Soil. 2010; 328(1-2):95-108. doi: 10.1007/s11104-009-0085-z

[14] Agam N, Berliner PR. Dew formation and water vapor adsorption in semi-arid environments - A review. J Arid Environ. 2006; 65(4):572-90. doi: 10. 1016/j.jaridenv.

2005. 09. 004

[15] Laio F, Porporato A, Ridolfi L, Rodriguez-Iturbe I. Plants in water-controlled ecosystems: active role in hydrologic processes and response to water stress - II.

Probabilistic soil moisture dynamics. Adv Water Resour. 2001; 24(7):707-23. doi:

10.1016/S0309-1708(01)00005-7

[16] D'Odorico P, Caylor K, Okin GS, Scanlon TM. On soil moisture-vegetation feedbacks and their possible effects on the dynamics of dryland ecosystems. J Geophys Res-Biogeo.

2007; 112(G4). doi: 10.1029/2006JG000379

[17] Wang LX, D'Odorico P, Manzoni S, Porporato A, Macko S. Soil carbon and nitrogen dynamics in southern African savannas: the effect of vegetation-induced patch-scale heterogeneities and large scale rainfall gradients. Climatic Change. 2009; 94(1-2):63-76. doi: 10.1007/s10584-009-9548-8

[18] D'Odorico P, Laio F, Porporato A, Rodriguez-Iturbe I. Hydrologic controls on soil carbon and nitrogen cycles. II. A case study. Adv Water Resour. 2003; 26(1):59-70. doi:

10.1016/S0309-1708(02)00095-7

[19] Stone AEC. Age and dynamics of the Namib Sand Sea: A review of chronological evidence and possible landscape development models. J Afr Earth Sci. 2013; 82:70-87. doi:

10. 1016/j.jafrearsci.2013.02.003

[20] Hachfeld B, Jurgens N. Climate patterns and their impact on the vegetation in a fog driven desert: The Central Namib Desert in Namibia. Phytocoenologia. 2000; 30(3-4):567-

89. doi: 10.1127/phytp/30/2000/567

[21] Lancaster J, Lancaster N, Seely M. Climate of the central Namib Desert. MADOQUA.

1984; 14(1):5-61.

[22] Eckardt FD, Soderberg K, Coop LJ, Muller AA, Vickery KJ, Grandin RD, et al. The nature of moisture at Gobabeb, in the central Namib Desert. J Arid Environ. 2013; 93:7-

19. doi: 10.1016/j.jaridenv.2012.01.011

[23] Kaseke KF, Wang LX, Wanke H, Turewicz V, Koeniger P. An analysis of precipitation isotope distributions across Namibia using historical data. PLoS One. 2016;

11(5). doi: 10.1371/journal.pone.0154598

[24] Kimura K. Origin of the fog in Namib Desert in dry season. Afr Study Monogr. 2005;

30:57-64.

[25] Kaseke KF, Mills AJ, Esler K, Henschel J, Seely MK, Brown R. Spatial variation of

"non-rainfall" water input and the effect of mechanical soil crusts on input and evaporation.

Pure Appl Geophys. 2012; 169(12):2217-29. doi:10.1007/s00024-012-0469-5

[26] Henschel JR, Seely MK. Ecophysiology of atmospheric moisture in the Namib Desert.

Atmos Res. 2008; 87(3-4):362-8. doi:10.1016/j.atmosres.2007.11.015

[27] Wang LX, Kaseke KF, Seely MK. Effects of non-rainfall water inputs on ecosystem functions. WIREs Water. 2016. In press. doi: 10.1002/wat1002.1179.

[28] Seely M. Irregular fog as a water source for desert dune beetles. Oecologia. 1979;

42(2):213-27. doi:10.1007/BF00344858

[29] Roth-Nebelsick A, Ebner M, Miranda T, Gottschalk V, Voigt D, Gorb S, et al. Leaf surface structures enable the endemic Namib desert grass Stipagrostis sabulicola to irrigate itself with fog water. J R Soc Interface. 2012; 9(73):1965-74. doi: 10.1098/rsif.2011.0847

[30] Hiatt C, Fernandez D, Potter C. Measurements of fog water deposition on the

California Central Coast. Atmospheric and Climate Sciences. 2012; 2(04):525. doi:

10.4236/acs/2012.24047

[31] Fischer DT, Still CJ, Williams AP. Significance of summer fog and overcast for drought stress and ecological functioning of coastal California endemic plant species. J

Biogeogr. 2009; 36(4):783-99. doi: 10.1111/j.1365-2699.2008.02025.x

[32] Bristow CS, Duller GAT, Lancaster N. Age and dynamics of linear dunes in the Namib

Desert. Geology. 2007; 35(6):555-8. doi:10.1130/G23369A.1

[33] Scholz H. The soils of the central Namib Desert with special consideration of the soils in the vicinity of Gobabeb. MADOQUA. 1972; 1:33-51.

[34] Jacobson K, van Diepeningen A, Evans S, Fritts R, Gemmel P, Marsho C, et al. Non- rainfall moisture activates fungal decomposition of surface litter in the Namib Sand Sea.

PLoS One. 2015; 10(5). doi: 10.1371/journal.pone.0126977

[35] Berndtsson R, Nodomi K, Yasuda H, Persson T, Chen HS, Jinno K. Soil water and temperature patterns in an arid desert dune sand. J Hydrol. 1996; 185(1-4):221-40. doi:10.1016/0022-1694(95)02987-7

[36] Berndtsson R, Chen HS. Variability of soil-water content along a transect in a desert area. J Arid Environ. 1994; 27(2):127-39. doi:10.1006/jare.1994.1053

[37] Fan JL, Scheuermann A, Guyot A, Baumgartl T, Lockington DA. Quantifying spatiotemporal dynamics of root-zone soil water in a mixed forest on subtropical coastal sand dune using surface ERT and spatial TDR. J Hydrol. 2015; 523:475-88. doi:

10.1016/j.jhydrol.2015.01.064

[38] Paige GB, Keefer TO. Comparison of field performance of multiple soil moisture sensors in a semi-arid rangeland. J Am Water Resour As. 2008; 44(1):121-35. doi:

10.1111/j.1752-1688.2007.00142.x

[39] Hu W, Shao MA, Han FP, Reichardt K, Tan J. Watershed scale temporal stability of soil water content. Geoderma. 2010; 158(3-4):181-98. doi:

10.1016/j.geoderma.2010.04.030

[40] Albergel C, de Rosnay P, Gruhier C, Munoz-Sabater J, Hasenauer S, Isaksen L, et al.

Evaluation of remotely sensed and modelled soil moisture products using global ground- based in situ observations. Remote Sens Environ. 2012; 118:215-26. doi:

10.1016/j.rse.2011.11.017

[41] Pan FF, Nieswiadomy M, Qian S. Application of a soil moisture diagnostic equation for estimating root-zone soil moisture in arid and semi-arid regions. J Hydrol. 2015;

524:296-310. doi: 10.1016/j.jhydrol.2015.02.044

[42] Devitt DA, Smith SD. Root channel macropores enhance downward movement of water in a Mojave Desert ecosystem. J Arid Environ. 2002; 50(1):99-108. doi:10.1006/jare.2001.0853

[43] Alfieri L, Claps P, D'Odorico P, Laio F, Over TM. An analysis of the soil moisture feedback on convective and stratiform precipitation. J Hydrometeorol. 2008; 9(2):280-91. doi: 10.1175/2007JHM863.1

[44] Grayson RB, Western AW, Chiew FHS, Bloschl G. Preferred states in spatial soil moisture patterns: Local and nonlocal controls. Water Resour Res. 1997; 33(12):2897-908. doi: 10.1029/97WR02174

[45] Gao L, Shao MG. Temporal stability of soil water storage in diverse soil layers. Catena.

2012; 95:24-32. doi: 10.1016/j.catena.2012.02.020

[46] Zhang X, Zhao W, Liu Y, Fang X, Feng Q. The relationships between grasslands and soil moisture on the Loess Plateau of China: A review. Catena. 2016; 145:56-67. doi:

10.1016/j.catena.2016.05.022.

[47] Hennessy J, Gibbens R, Tromble J, Cardenas M. Water properties of caliche. Journal of Range Management. 1983:723-6.

[48] Yamagata K, Mizuno K. Landform development along the middle course of the Kuiseb

River in the Namib Desert, Namibia. 2005.

[49] Laio F. A vertically extended stochastic model of soil moisture in the root zone. Water

Resour Res. 2006; 42(2). doi: 10.1029/2005WR004502

Chapter 3: The impact of non – rainfall on soil moisture dynamics in the Namib

Desert

3.1 Introduction

Drylands cover 40% of the earth surface, and are characterized by regions where mean annual precipitation is significantly lower than potential evapotranspiration (PET) [1, 2].

Drylands are critical systems inhabited by about 38% of the global population [3, 4], of which 90% of live in developing countries [2]. Drylands are also home to a significant number of flora and fauna, contribute to 40% of the global net primary productivity (NPP) and account for over one third of the global carbon stock in the form of soil carbon [5, 6].

Because of the close linkage between vegetation dynamics and dryland soil moisture, soil moisture is critical in maintaining the functionality of dryland ecosystems [7, 8]. Spatial heterogeneity of root zone soil moisture was reported to be one of the primary contributors to the formation of vegetation patterns in some dryland ecosystems [9-11]. For example in central Kenya, the formation and expansion of a two-phase pattern of Sansevieria volkensii is due to “soil moisture halo effect” [12]. While tree-grass coexistence patterns in the

Kalahari Desert is primarily induced by differences in soil water balance and plant water stress [13]. Differences in soil moisture were reported as one of the main reasons for low seedling establishment observed under inter-canopy versus canopy environments [14, 15].

In addition, some abiotic factors and physical processes are affected by soil moisture. For instance, a twenty years projection (2080-2099) from multiple modeling results suggests that global surface soil moisture to drop by 5 to 15%, which may indirectly influence soil organic carbon stock and total nitrogen in drylands [16, 17]. Land-surface interactions can also be influenced by soil moisture since the presence of soil moisture darkens surface soil

resulting in the changing of surface albedo and air temperature, which may significantly alter near surface climate [18-20].

Defined as suspended water drops in the atmosphere near the Earth’s surface, fog is an important supplementary water source for human utilization, sustaining the survival of flora and fauna and maintaining biogeochemical cycling [21, 22]. Previous investigations demonstrated that fog comprised a significant amount of the annual hydrologic input of

California redwood forest [23, 24]. A three-year investigation showed that up to 19 % of water used by redwood trees originated from fog during the dry summer season, while up to 66% of water of understory plants was from fog [25]. This phenomenon is much more important in the drylands where water is limiting and fog amount may exceed annual rainfall [26]. The unique leaf structure and physiology of an endemic Namib Desert grass,

Stipagrostis sabulicola, make it an efficient fog harvester transferring fog water to the plant base by means of stemflow and it is thus heavily reliant on fog water [27]. By means of a spray experiment, researchers concluded that another Namib species, Trianthema hereorensis, was able to survive in the southern Namib dune system by distributing leaf- absorbed fog water to the rest part of the plant [28]. Similar results were also found in other drylands. To investigate why dwarf succulents were able to survive in an arid environment of South Africa with poor leaf and stem development, comparisons of atmospheric moisture interception by gravel and two dwarf succulents (Agyroderma pearsonii and

Cepphalophyllum spissum) indicated that fog absorption contributes nearly half of the total water absorbed by those two dwarf succulents [29]. The results indicate that fog is as vital as rainfall in sustaining growth and survival of dwarf succulents in these arid environments.

Although some dryland studies have highlighted the role of soil moisture and fog on maintaining plant development and biogeochemical processes, little is known about the relationship between fog and soil moisture dynamics. To our knowledge, only few snapshot observations have been made between fog and soil moisture dynamics. For example, observations in California coastal pine forest showed that fog was an important contributor to the re-wetting of soil during rain-free periods [23, 30]. However, no previous studies have investigated how and to what extent fog can affect soil moisture dynamics. Moreover, most dryland fog observations concentrate on the effect of fog on vegetation water status than on soil biogeochemical processes. There is still a lack of characterization of fog statistical distribution particularly in the Namib Desert where fog is a frequent occurrence

[31]. Characterizing the distribution of fog distribution is an important step toward quantitatively describing fog dynamics and predicting the changes in fog patterns.

Therefore, to address these knowledge gaps, the objectives of this study were to, 1) quantify the statistical distribution of fog; 2) fill in data gaps of fog and soil moisture dynamics in the Namib Desert; 3) modify a stochastic modeling framework to simulate the effects of fog on soil moisture dynamics during rainless periods.

3.2 Material and Methods

3.2.1 Site descriptions

The field observations were conducted at three locations (gravel plains at Gobabeb, here after GPG; sand dunes at Gobabeb, here after SDG and gravel plains at Kleinberg, here after GPK) from two sites (Gobabeb and Kleinberg) within the Namib Desert.

Gobabeb (lat. - 23.55° S, long. 15.04° E, and elv. 405 m a.s.l) is located 60 km from the

Atlantic Ocean south-east of Walvis Bay on the banks of the Kuiseb River and at the edge of the Namib Sand Sea [32] (Fig. 3.1a & Fig. 3.1b). The climate is hyper-arid and the frequency of rainfall is extremely low with a mean annual rainfall of 27 mm [33]. Wet season and dry season of Gobabeb are pronounced with December to May being the rainy season and June to November being the dry season. The mean annual temperature of

Gobabeb is 21.1℃ (mean monthly temperature ranging from 17.7 to 24.2℃) [34, 35]. The average relative humidity of Gobabeb is around 50% with most of the moisture derived from fog [35]. The mean annual foggy days at Gobabeb is ninety-four days, which is nearly fifty days less than that of Walvis Bay where fog is strongly influenced by the cold

Benguela current [36]. The ephemeral Kuiseb River separates the Sand Sea and the gravel plains (gypcrete) north and south of Gobabeb, respectively [37] (Fig. 3.1a and Fig. 3.1b).

The dominant plant species in the gravel plains are Zygophyllum simplex and Z. stapffi while Stipagrostis sabulicola and Trianthema heroensis are the dominant species in the sand dune area [38].

Kleinberg (lat. -22.98° S, long. 14.73° E and elv. 180 m a.s.l) is located 33 km the

Atlantic Ocean and has been a Gobabeb Research and Training Centre (GRTC) field site since 1982 [39]. The mean annual temperature is generally lower at Kleinberg than that of

Gobabeb where July being the coldest month compared with September at Kleinberg [35].

Most areas of Kleinberg are dominated by gravel plains (Fig. 3.1c) with high salinity and low organic matter inhabited by pencil bush (Arthraerua leubnitziae) and lichen fields.

3.2.2 Data collection

Soil moisture was measured at hourly intervals using the CS655 Water Content

Reflectometer (Campbell Scientific, Inc. Logan, Utah, USA) from three locations. At GPG

a single probe was installed under bare soil at 4 cm depth. At SDG two probes were installed at 4 cm soil depth, one under bare soil and the other under vegetation. At GPK one probe was installed under bare soil at 5 cm depth. All soil moisture probes were installed horizontally at the field sites. Fog data was obtained from FogNet stations (Fig.

3.1d), which are part of the Southern African Science Service Centre from Climate Change and Adaptive Land Management (SASSCAL). Each FogNet station comprised a cylindrical passive fog collector (Juvik fog collector) coupled with regular rainfall gauge and screen mesh to measure fog amount every second. Due to the close proximity of SDG and GPG (approximate 3.5 km apart), data from the same fog collector was used. Soil saturated hydraulic conductivity (Ks) at each site was derived from literature [40]. Eighty rainless days (August 19, 2015 to November 6, 2015) continuous volumetric soil moisture data and approximately two years’ (December 1, 2014 to November 1, 2016) fog data were used for field data analysis and modeling purposes.

3.2.3 Analyses of field data

In order to evaluate the effect of fog on soil moisture dynamics, hourly soil moisture data and fog data were processed to daily scale. Central tendency and variability of fog and soil moisture data were expressed as mean, standard deviation (S.D.) and coefficient of variation (CV). To characterize the distribution of fog, a graphic method was used by visually examining histograms of field fog data and a sequence of data generated by a non- homogeneous Poisson process [41].

3.2.4 Model structure

A process-based modeling framework was used to simulate soil moisture dynamics with fog as the sole water input variable. The model was originally developed to understand how the stochastic rainfall influences soil moisture dynamics in drylands by expressing rainfall as a non-homogeneous Poisson process [42]. In this study, the model was modified by replacing rainfall with fog water input and used a deterministic approach using field fog data to drive the model. A simplified stochastic differential equation for bare soil water balance over the layer of depth Zr is expressed as follows:

ds n푍 = µ퐹 − 퐸(푠) − 푇(푠) − 퐿(푠), (1) 푟 dt where n is soil porosity, Zr is the active soil depth, s is relative soil moisture, µ is a fog parameter that represents the percentage of fog intercepted by soil surface, F is the amount of fog collected by fog collector, E(s) and T(s) are moisture loss through evaporation and transpiration respectively, L(s) is leakage via the bottom layer. In this study, T(s) was set to zero because the modeling was applied to a bare soil ground. The modified model assumes that all fog water deposited on the soil surface is immediately transported into the soil

(infiltration) and no leakage or surface runoff is generated (i.e., L(s) = 0).

Therefore according to this modified framework, the increase in soil moisture is due to fog infiltration. The loss of soil moisture is only due to soil evaporation. The loss function can be expressed as:

0 0 < 푠 ≤ 푠ℎ

퐸(푠) = 푠 −푠ℎ , (2) 퐸vap ∗ 푠ℎ < 푠 ≤ 푠w 푠푤−푠h

{ 퐸vap 푠w < 푠 ≤ 1

where s is the relative soil moisture, sh is soil moisture at the hygroscopic point, sw is soil moisture at the wilting point, Evap is the soil evaporation. For s > sw, evaporation will reach its maximum rate. For sh< s ≤ sw, soil moisture starts to restrict evaporation and a positive

relationship between soil moisture and evaporation is found. For s ≤ sh, no evaporation is generated. Details of modeling parameters that were used in this study can be found in

Table 3.1.

Table 3.1 Soil and fog parameters for gravel plain (Gobabeb). Gravel plain (Gobabeb) Vegetation coverage Bare soil Soil parameters Porosity†, n 0.34 † Field capacity , sfc 0.1 * Hydroscopic point , sh 0.04 † -1 Saturated hydraulic conductivity , Ks (m day ) 5.6 * Soil depth , Zr (m) 0.34 * -1 Evaporation , Evap (mm day ) 0.65 Fog parameter Fog arrival rate*, λ (day-1) 0.3 Average fog depth*, α (mm) 1.51 Fog absorption factor*, µ 0.13 Soil - vegetation parameter * Permanent wilting point , sw 0.085 †Li et al. (2016) *This study

3.3 Results and Discussion

3.3.1 Fog distribution

In the recent stochastic soil moisture modeling framework, rainfall was assumed to be a Poisson process with a rate parameter λ and each event carried a random amount of

rainfall α, which follow an exponential distribution [41, 43]. This is in coordinate with the occurrence of fog because the occurrence of each fog event is independent and each fog event carries a random amount of water. This suggests that fog and rainfall potentially share a similar distribution. We derived λ and α parameters using two year fog field observations from the Gobabeb and Kleinberg FogNet stations (Table 3.2).

Table 3.2 Soil vegetation coverage, soil and fog parameters at three field sites spanning from August 19, 2015 to November 6, 2015. Note: No soil moisture dynamics were observed in GPK during the study period. Field sites Depth (cm) MSM (%) CV (%) Total fog (mm) λ (day-1) α (mm) GPG 4 3.3 Bare soil 1.54±0.05 SDG 4 19.6 Bare soil 0.51±0.1 36.2 0.3 1.51 SDG 4 18.8 Vegetated 0.53±0.1 GPK 5 0 0 89.8 0.55 2.04 Bare soil

A sequence of data was generated from a Poisson process using the two derived parameters.

By plotting field observed fog against derived data set at Gobabeb and Kleinberg, the results showed that histograms between field observed fog and derived fog in these two locations generally showed a similar pattern suggesting that the two groups of data can be generated from the distribution (Fig. 3.2a and Fig. 3.2b). This suggests that fog can be characterized using a Poisson process. Unveiling fog distribution particularly in arid regions is of great value. For example, fog waters in some drylands were reported to include a substantial amount of elements and were clean enough for human drinking and production purposes [44, 45]. A better understanding of distribution of fog deposition may enhance the rationality of when and where to install the fog harvesting systems, which could dramatically improve the efficiency of fog harvest. Vegetation patterns were also

found to have close links with fog deposition because vegetation not only benefits from fog water (moisture) but also the various essential nutrients from fog water for growth [46,

47]. By characterizing fog distribution and incorporating this into ecohydrological models, it becomes feasible to project changes in vegetation dynamics induced by fog pattern changes under the context of global climate change.

3.3.2 Field observations

Table 2 shows fog parameters based on eighty day field observations (August 19, 2015 to November 6, 2015) recorded from the three locations. Total fog amount for GPK was

89.8 mm (Table 3.2), which was significantly higher than that of GPG and SDG (36.2 mm,

Table 3.2). The frequency (λ) and average depth (α) of fog exhibited different patterns at these two locations, with a larger average fog depth and more foggy days occurring at GPK.

The differences between the fog total amount and fog parameters at Gobabeb and Kleinberg may be affected by the elevation, topography and location (e.g., distance to the ocean) of fog gauges.

Figure 3.2 Comparison of histograms between field observed fog and Poisson simulated fog at Gobabeb (a) and Kleinberg (b).

Fig. 3.3 shows soil moisture dynamics and its relationships with fog events at three study sites. The mean soil moisture at GPG was 1.55%, which is approximately three times higher than that at SDG (0.51% under bare soil, 0.53% under vegetated soil, respectively,

Table 2) regardless of vegetation cover. Their CVs differed with CV at SDG under bare soil being the largest (19.6%, Table 3.2), which is nearly six times more than that of GPG

(P < 0.05). The differences between mean soil moisture among three study sites might be explained by their differences in soil texture [40]. The soil moisture observations at SDG suggests that the mean soil moisture for vegetated soil (0.53%, Table 3.2) were slightly higher than that of bare soil (0.51%, Table 3.2).

During the rainless period, fog was observed to have moderate impacts on soil moisture dynamics with the rising of soil moisture correspondent to a series of fog events at GPG (Fig. 3.3a). At SDG, the relationship between soil moisture and fog tended to be weaker though some soil moisture peaks matched with fog events (Fig. 3.3b). A clear

discrepancy in soil moisture dynamics between SDG and GPG (Fig. 3.3a, Fig. 3.3b) existed under the same fog regimes. The discrepancy might be due to the differences in soil texture since gravel plain has a stronger water hold capacity than sand dune when given the same water input. At GPK, no soil moisture dynamics (Fig. 3.3c) were observed to be related to fog occurrences, which may be attributed to the presence of soil crusts on the soil surface at GPK. They might act as an impermeable layer impeding water infiltration, particularly preventing small amounts of water (e.g., fog, water vapor adsorption and dew) to be absorbed by the soil surface [39].

Figure 3.3 Fog events and soil moisture dynamics at sand dune (Gobabeb), gravel plain (Kleinberg) and gravel plain (Gobabeb).

In summary, soil moisture dynamics was observed to have moderate correlation with fog events at GPG. A weak relationship between fog and soil moisture were found at SDG and there was no relationship when moving further west to GPK during the rainless period.

During the course of wet periods (e.g., during rainy reason), no soil moisture and fog relationships were found at any of those three sites (data not shown). This is because the occurrences of rainfall events were mainly concentrated in the summer season. Even a small amount of rainfall may affect soil moisture dynamics for a long time and might mask the effect of fog on soil moisture dynamics.

Our field observations filled the data gaps in concurrent fog and soil moisture observations in the Namib Desert and provided data support for studying vegetation and animal adaptions in the fog dependent systems. In addition, predictions in this already arid desert indicated that there would be less rainfall or larger rainfall variability in the future

[48]. Knowledge of the soil moisture-fog relationship during the rainless periods suggested that stochastic modeling frameworks coupled with fog parameters can be used for future soil moisture predictions.

3.3.3 Soil moisture modeling with fog as the sole water input

Soil moisture dynamics at GPG during rainless periods was selected and simulated by a modified stochastic modeling framework driven by field fog observations. In general, overall soil moisture patterns can be captured using this modeling framework and mean relative soil moisture estimated from the modeling agreed well with field observations. The simulated mean relative soil moisture was 0.0442, which was close to that observed in the field (0.0454) (Fig. 3.4). Most of the simulated soil moisture peaks and observed soil

moisture peaks matched (Fig. 3.4). This implied the feasibility of the modified modeling framework for future projections. Such a modeling framework would be particularly useful for drylands such as the Namib Desert where rainfall is rare but fog is frequent.

Figure 3.4 Comparison between field soil moisture dynamics and simulated soil moisture dynamics at gravel plain (Gobabeb, GPG) at the depth of 4 cm.

Although the overall soil moisture dynamics can be simulated, some modeled moisture peaks did not match the field observations (Fig. 3.4). The mismatch between field soil moisture peaks and simulated peaks may be affected by how the amount of fog is estimated. Fog is suspended water droplets and it forms only when the atmosphere water vapor reaches saturation [21]. At field sites, fog collectors are installed above the ground

[44]. Because of this arrangement, fog water collected by fog collectors is not necessarily the actual fog that deposited on the soil surface, which might one of the reasons why there are mismatch between simulated soil moisture peaks and observed peaks. In addition, the infiltration mechanism of fog water is still poorly understood. For example, a heavy fog event doesn’t mean more fog infiltration to the soil profile and in turn a small fog event unnecessarily indicates less fog infiltration. In addition, moisture input into soil may start

earlier as water vapor adsorption, which the modified framework failed to take into account.

These three uncertainties may be responsible for the mismatch of soil moisture peaks.

Moreover, although soil moisture dynamics can be simulated using fog as a sole water input during rainless periods, wet season soil moisture dynamics may not be fully revealed by the modeling framework, which requires further improvements toward a better understanding of fog characterization, fog infiltration and fog-soil moisture relationships.

3.4 Conclusions

In this study, we demonstrated that fog can be well-characterized by a non- homogeneous Poisson process with two parameters (fog arrival rate and average depth).

Our fog distribution investigation provided new insights and modeling support for future ecohydrological studies. For example, fog influenced vegetation dynamics in drylands can be predicted by coupling ecohydrological models with fog parameters. Soil moisture and fog analyses from three field sites within the Namib Desert suggested that soil moisture dynamics were affected by fog occurrence at GPG, while the relationship became less pronounced at SDG and there was no relationship at GPK. The field results and analyses filled the concurrent fog and soil moisture observation data gap in the Namib Desert and shed light on using ecohydrological models to couple fog parameter with soil moisture dynamics. Informed by field observations, a stochastic modeling framework was used to simulate the impact of fog on soil moisture dynamics. The modeling results showed that most of soil moisture peaks and mean relative soil moisture were well captured by the modeling framework. This suggests the feasibility of using this modified framework to predict future soil moisture changes under changing fog conditions. However, the fog

impact on soil moisture during the rainy season cannot be captured due to residual effect of rainfall that may mask the impact of fog on soil moisture dynamics, which might require future works.

References

[1] Maestre FT, DJ Eldridge, S Soliveres, S Kefi, M Delgado-Baquerizo, MA Bowker, et al. Structure and functioning of dryland ecosystems in a changing world. Annu Rev Ecol

Evol Syst. 47 (2016) 215-37, doi: 10.1146/annurev-ecolsys-121415-032311.

[2] Wang L, P D'Odorico, JP Evans, DJ Eldridge, MF McCabe, KK Caylor, et al. Dryland ecohydrology and climate change: critical issues and technical advances. Hydrology and

Earth System Sciences. 16 (2012) 2585-603, doi: 10.5194/hess-16-2585-2012.

[3] Maestre FT, R Salguero-Gomez, JL Quero. It is getting hotter in here: determining and projecting the impacts of global environmental change on drylands. Philos Trans R Soc

Lond B Biol Sci. 367 (2012) 3062-75, doi: 10.1098/rstb.2011.0323.

[4] Schimel DS. Climate. Drylands in the Earth system. Science. 327 (2010) 418-9, doi:

10.1126/science.1184946.

[5] Plaza-Bonilla D, JL Arrúe, C Cantero-Martínez, R Fanlo, A Iglesias, J Álvaro-Fuentes.

Carbon management in dryland agricultural systems. A review. Agronomy for Sustainable

Development. 35 (2015) 1319-34, doi: 10.1007/s13593-015-0326-x.

[6] Zika M, KH Erb. The global loss of net primary production resulting from human- induced soil degradation in drylands. Ecological Economics. 69 (2009) 310-8, doi:

10.1016/j.ecolecon.2009.06.014.

[7] Yu K, P D'Odorico. An ecohydrological framework for grass displacement by woody plants in savannas. Journal of Geophysical Research: Biogeosciences. 119 (2014) 192-206, doi: 10.1002/2013jg002577.

[8] Zhang S, L Lovdahl, H Grip, Y Tong, X Yang, Q Wang. Effects of mulching and catch cropping on soil temperature, soil moisture and wheat yield on the Loess Plateau of China.

Soil and Tillage Research. 102 (2009) 78-86, doi: 10.1016/j.still.2008.07.019.

[9] Ravi S, P D'Odorico, L Wang, S Collins. Form and function of grass ring patterns in arid grasslands: the role of abiotic controls. Oecologia. 158 (2008) 545-55, doi:

10.1007/s00442-008-1164-1.

[10] Ruiz-Sinoga JD, JF Martínez-Murillo, MA Gabarrón-Galeote, R García-Marín. The effects of soil moisture variability on the vegetation pattern in Mediterranean abandoned fields (Southern Spain). Catena. 85 (2011) 1-11, doi: 10.1016/j.catena.2010.11.004.

[11] Zribi M, TP Anguela, B Duchemin, Z Lili, W Wagner, S Hasenauer, et al. Relationship between soil moisture and vegetation in the Kairouan plain region of Tunisia using low spatial resolution satellite data. Water Resources Research. 46 (2010), doi: Artn W06508

10.1029/2009wr008196.

[12] Franz TE, EG King, KK Caylor, DA Robinson. Coupling vegetation organization patterns to soil resource heterogeneity in a central Kenyan dryland using geophysical imagery. Water Resources Research. 47 (2011), doi: 10.1029/2010wr010127.

[13] Porporato A, F Laio, L Ridolfi, KK Caylor, I Rodriguez-Iturbe. Soil moisture and plant stress dynamics along the Kalahari precipitation gradient. J Geophys Res-Atmos. 108

(2003), doi: 10.1029/2002jd002448.

[14] Casper BB, HJ Schenk, RB Jackson. Defining a plant's belowground zone of influence.

Ecology. 84 (2003) 2313-21, doi: 10.1890/02-0287.

[15] Hamerlynck EP, RL Scott, M Susan Moran, AM Schwander, E Connor, TE Huxman.

Inter- and under-canopy soil water, leaf-level and whole-plant gas exchange dynamics of a semi-arid perennial C4 grass. Oecologia. 165 (2011) 17-29, doi: 10.1007/s00442-010-

1757-3.

[16] Dai A. Increasing drought under global warming in observations and models. Nature

Climate Change. 3 (2012) 52-8, doi: 10.1038/nclimate1633.

[17] Huang J, H Yu, X Guan, G Wang, R Guo. Accelerated dryland expansion under climate change. Nature Climate Change. (2015), doi: 10.1038/nclimate2837.

[18] Guan X, J Huang, N Guo, J Bi, G Wang. Variability of soil moisture and its relationship with surface albedo and soil thermal parameters over the Loess Plateau. Advances in

Atmospheric Sciences. 26 (2009) 692-700, doi: 10.1007/s00376-009-8198-0.

[19] Liu H, B Wang, C Fu. Relationships between surface albedo, soil thermal parameters and soil moisture in the semi-arid area of Tongyu, northeastern China. Advances in

Atmospheric Sciences. 25 (2008) 757-64, doi: 10.1007/s00376-008-0757-2.

[20] Rutherford WA, TH Painter, S Ferrenberg, J Belnap, GS Okin, C Flagg, et al. Albedo feedbacks to future climate via climate change impacts on dryland biocrusts. Sci Rep. 7

(2017) 44188, doi: 10.1038/srep44188.

[21] Wang L, KF Kaseke, MK Seely. Effects of non-rainfall water inputs on ecosystem functions. Wiley Interdisciplinary Reviews: Water. 4 (2017) e1179, doi:

10.1002/wat2.1179.

[22] Fu G, JT Guo, SP Xie, YH Duane, MG Zhang. Analysis and high-resolution modeling of a dense sea fog event over the Yellow Sea. Atmospheric Research. 81 (2006) 293-303, doi: 10.1016/j.atmosres.2006.01.005.

[23] Fischer DT, CJ Still, CM Ebert, SA Baguskas, AP Williams. Fog drip maintains dry season ecological function in a California coastal pine forest. Ecosphere. 7 (2016) e01364, doi: 10.1002/ecs2.1364.

[24] Limm EB, KA Simonin, AG Bothman, TE Dawson. Foliar water uptake: a common water acquisition strategy for plants of the redwood forest. Oecologia. 161 (2009) 449-59, doi: 10.1007/s00442-009-1400-3.

[25] Dawson TE. Fog in the California redwood forest: ecosystem inputs and use by plants.

Oecologia. 117 (1998) 476-85, doi: 10.1007/s004420050683.

[26] Agam N, PR Berliner. Dew formation and water vapor adsorption in semi-arid environments - A review. Journal of Arid Environments. 65 (2006) 572-90, doi:

10.1016/j.jaridenv.2005.09.004.

[27] Roth-Nebelsick A, M Ebner, T Miranda, V Gottschalk, D Voigt, S Gorb, et al. Leaf surface structures enable the endemic Namib desert grass Stipagrostis sabulicola to irrigate itself with fog water. J R Soc Interface. 9 (2012) 1965-74, doi: 10.1098/rsif.2011.0847.

[28] Seely MK, MPd Vos, GN Louw. Fog imbibition satellite fauna and unusual leaf structure in a Namib. South African Journal of Science. 73 (1977) 169-72.

[29] Matimati I, CF Musil, L Raitt, E February. Non rainfall moisture interception by dwarf succulents and their relative abundance in an inland arid South African ecosystem.

Ecohydrology. (2012) 818-25, doi: 10.1002/eco.1304.

[30] Ewing HA, KC Weathers, PH Templer, TE Dawson, MK Firestone, AM Elliott, et al.

Fog Water and Ecosystem Function: Heterogeneity in a California Redwood Forest.

Ecosystems. 12 (2009) 417-33, doi: 10.1007/s10021-009-9232-x.

[31] Olivier J. Spatial-Distribution of Fog in the Namib. Journal of Arid Environments. 29

(1995) 129-38, doi: 10.1016/S0140-1963(05)80084-9.

[32] Kaseke KF, L Wang, MK Seely. Nonrainfall water origins and formation mechanisms.

Sci Adv. 3 (2017) e1603131, doi: 10.1126/sciadv.1603131.

[33] Eckardt FD, K Soderberg, LJ Coop, AA Muller, KJ Vickery, RD Grandin, et al. The nature of moisture at Gobabeb, in the central Namib Desert. Journal of Arid Environments.

93 (2013) 7-19, doi: 10.1016/j.jaridenv.2012.01.011.

[34] Goudie A. Climate, weathering, crust formation, dunes, and fluvial features of the

Central Namib Desert, near Gobabeb, South West Africa. MADOQUA. 2 (1970) 15-31.

[35] Lancaster J, N Lancaster. climate of the central namib desert. MADOQUA. 14 (1984)

5-61.

[36] Cermak J. Low clouds and fog along the South-Western African coast - Satellite-based retrieval and spatial patterns. Atmospheric Research. 116 (2012) 15-21, doi:

10.1016/j.atmosres.2011.02.012.

[37] Gottsche FM, FS Olesen, A Bork-Unkelbach. Validation of land surface temperature derived from MSG/SEVIRI with in situ measurements at Gobabeb, Namibia. Int J Remote

Sens. 34 (2013) 3069-83, doi: 10.1080/01431161.2012.716539.

[38] Jacobson K, A van Diepeningen, S Evans, R Fritts, P Gemmel, C Marsho, et al. Non- rainfall moisture activates fungal decomposition of surface litter in the Namib Sand Sea.

PLoS One. 10 (2015) e0126977, doi: 10.1371/journal.pone.0126977.

[39] Kaseke KF, AJ Mills, K Esler, J Henschel, MK Seely, R Brown. Spatial Variation of

“Non-Rainfall” Water Input and the Effect of Mechanical Soil Crusts on Input and

Evaporation. Pure and Applied Geophysics. 169 (2012) 2217-29, doi: 10.1007/s00024-

012-0469-5.

[40] Li B, L Wang, KF Kaseke, L Li, MK Seely. The Impact of Rainfall on Soil Moisture

Dynamics in a Foggy Desert. PLoS One. 11 (2016) e0164982, doi:

10.1371/journal.pone.0164982.

[41] Rodriguez-Iturbe I, A Porporato, L Ridolfi, V Isham, DR Cox. Probabilistic modelling of water balance at a point: the role of climate, soil and vegetation. P Roy Soc a-Math Phy.

455 (1999) 3789-805.

[42] Laio F, A Porporato, L Ridolfi, I Rodriguez-Iturbe. Plants in water-controlled ecosystems: active role in hydrologic processes and response to water stress - II.

Probabilistic soil moisture dynamics. Advances in Water Resources. 24 (2001) 707-23, doi:

10.1016/S0309-1708(01)00005-7.

[43] Porporato A, P D'Odorico, F Laio, I Rodriguez-Iturbe. Hydrologic controls on soil carbon and nitrogen cycles. I. Modeling scheme. Advances in Water Resources. 26 (2003)

45-58, doi: 10.1016/S0309-1708(02)00094-5.

[44] Klemm O, RS Schemenauer, A Lummerich, P Cereceda, V Marzol, D Corell, et al.

Fog as a fresh-water resource: overview and perspectives. Ambio. 41 (2012) 221-34, doi:

10.1007/s13280-012-0247-8.

[45] Eckardt FD, RS Schemenauer. Fog water chemistry in the Namib Desert Namibia.

Atmospheric Environment. 32 (1998) 2595-9, doi: 10.1016/S1352-2310(97)00498-6.

[46] Beiderwieden E, T Wrzesinsky, O Klemm. Chemical characterization of fog and rain water collected at the eastern Andes cordillera. Hydrology and Earth System Sciences. 9

(2005) 185-91.

[47] Templer PH, KC Weathers, HA Ewing, TE Dawson, S Mambelli, AM Lindsey, et al.

Fog as a source of nitrogen for redwood trees: evidence from fluxes and stable isotopes.

Journal of Ecology. 103 (2015) 1397-407, doi: 10.1111/1365-2745.12462.

[48] Lu X, L Wang, M Pan, KF Kaseke, B Li. A multi-scale analysis of Namibian rainfall over the recent decade – comparing TMPA satellite estimates and ground observations.

Journal of Hydrology: Regional Studies. 8 (2016) 59-68, doi: 10.1016/j.ejrh.2016.07.003

Chapter 4: Conclusions

The first part of this thesis has filled a long-term soil moisture and rainfall data gap in the foggy Namib Desert. The field analyses suggested that soil moisture dynamics affected by rainfall at two gravel plain sites, while the relationship became less significant at the sand dune site. Our long-term field observations and field analyses provided the data support for understanding ecohydrological processes and offer a theoretical basis for soil moisture modeling in this fog dominant ecosystem. The modeling results generally in coordinate with field observations with rainfall events correspondent to soil moisture peaks.

Our modeling results suggested that the stochastic modeling can be used to project future soil moisture dynamics during rainy season in those two gravel plain sites. The modeling sensitivity in response to different soil texture yield consistent results among the study sites.

In general, soil moisture outputs were mainly controlled by parameters in bounded group

(e.g., soil hygroscopic point (sh) and field capacity (sfc)) and less sensitive to parameters in non-bounded group (e.g., saturated hydraulic conductivity (Ks) and soil porosity (n)). The sensitivity can be regard as an uncertainty indicator of this modeling framework for future applications.

Based on the field and modeling results, new field observations and a modified modeling framework were proposed in the second part of this thesis to evaluate the effects of fog on soil moisture dynamics. Fog distribution can be well-characterized by a non- homogeneous Poisson process with two parameters based on a two years’ fog observations.

This provide new insights and modeling support for other ecohydrological studies. The eighty days’ rainless periods’ soil moisture dynamics and the correspondent fog observations suggested that soil moisture dynamics were affected by fog events in GPG,

while this relationship becomes less pronounced in SDG. At GPK, no soil moisture dynamic and fog relationship were observed. The modified modeling framework in GPG indicated that soil moisture dynamics can be captured using fog as the solo water input during rainless periods. This makes it possible to apply this modified modeling framework to predict future soil moisture changes under the fog influence. Our field observations and modeling results are of great importance to understand and predict dryland soil moisture dynamics in the context of global climate change.

Chapter 5: Appendices

Appendix A: Soil moisture dynamics and rainfall patterns at Gobabeb.

Gravel plain Gobabeb Sand dune Gobabeb Time bare soil bare soil vegetation vegetation bare soil Rain Depth 7.5 cm 22.5 cm 7.5 cm 7.5 cm 15 cm Unit m^3/m^3 m^3/m^3 m^3/m^3 m^3/m^3 m^3/m^3 mm 1/2/2014 2.14 0.79 2.66 0.00 1/3/2014 2.11 0.84 2.70 0.00 1/4/2014 2.05 0.78 2.62 0.00 1/5/2014 2.10 0.78 2.72 0.00 1/6/2014 2.06 0.78 2.69 0.00 1/7/2014 2.03 0.77 2.65 0.00 1/8/2014 2.06 0.78 2.65 0.00 1/9/2014 2.08 0.79 2.65 0.00 1/10/2014 2.11 0.80 2.71 0.00 1/11/2014 2.12 0.83 2.79 0.00 1/12/2014 2.10 0.82 2.76 0.00 1/13/2014 2.12 0.83 2.84 0.00 1/14/2014 2.08 0.82 2.78 0.00 1/15/2014 2.11 0.83 2.83 0.00 1/16/2014 2.12 0.85 2.92 0.00 1/17/2014 2.12 0.85 2.92 0.00 1/18/2014 2.07 0.85 2.83 0.00 1/19/2014 2.12 0.80 2.92 0.00 1/20/2014 2.10 0.82 2.86 0.00 1/21/2014 2.09 0.82 2.91 0.00 1/22/2014 2.09 0.83 3.01 0.00 1/23/2014 2.08 0.81 2.99 0.00 1/24/2014 2.07 0.81 2.97 0.00 1/25/2014 2.07 0.81 2.94 0.00 1/26/2014 2.05 0.79 2.89 0.00 1/27/2014 2.02 0.76 2.83 0.00 1/28/2014 2.03 0.74 2.83 0.00 1/29/2014 2.05 0.75 2.87 0.00 1/30/2014 2.08 0.77 2.90 0.00 1/31/2014 2.05 0.77 2.90 0.00

2/1/2014 2.04 0.75 2.87 0.00 2/2/2014 2.05 0.73 2.89 0.00 2/3/2014 2.08 0.75 2.95 0.00 2/4/2014 2.03 0.74 2.90 0.00 2/5/2014 2.07 0.76 2.95 0.00 2/6/2014 2.07 0.76 2.93 0.00 2/7/2014 2.05 0.75 2.92 0.00 2/8/2014 2.05 0.75 2.92 0.00 2/9/2014 2.04 0.75 2.89 0.00 2/10/2014 2.05 0.73 2.86 0.00 2/11/2014 2.08 0.76 2.94 0.00 2/12/2014 2.04 0.74 2.93 0.00 2/13/2014 2.15 0.76 3.08 0.00 2/14/2014 2.15 0.82 3.10 0.00 2/15/2014 2.14 0.84 3.13 0.00 2/16/2014 2.15 0.86 3.16 0.00 2/17/2014 2.13 0.83 3.11 0.00 2/18/2014 2.15 0.83 3.13 0.00 2/19/2014 2.10 0.82 3.07 0.00 2/20/2014 2.07 0.79 3.06 0.00 2/21/2014 2.01 0.73 3.03 0.00 2/22/2014 2.00 0.72 3.00 0.00 2/23/2014 1.98 0.76 3.03 0.00 2/24/2014 1.97 0.79 3.02 0.00 2/25/2014 2.03 0.84 3.07 0.00 2/26/2014 1.97 0.81 2.98 0.00 2/27/2014 2.00 0.78 3.00 0.00 2/28/2014 2.00 0.79 3.03 0.00 3/1/2014 1.98 0.78 2.99 0.00 3/2/2014 1.93 0.75 2.92 0.00 3/3/2014 1.91 0.71 2.84 0.00 3/4/2014 1.93 0.71 2.83 0.00 3/5/2014 1.93 0.71 2.85 0.00 3/6/2014 1.95 0.73 2.86 0.00 3/7/2014 1.95 0.75 2.89 0.00 3/8/2014 1.95 0.75 2.83 0.00 3/9/2014 1.92 0.72 2.76 0.00 3/10/2014 1.89 0.69 2.70 0.00 3/11/2014 1.89 0.68 2.66 0.00 3/12/2014 1.87 0.66 2.65 0.00

3/13/2014 1.88 0.65 2.65 0.00 3/14/2014 1.90 0.68 2.70 0.00 3/15/2014 1.94 0.71 2.70 0.00 3/16/2014 1.98 0.77 2.76 0.00 3/17/2014 2.04 0.80 2.79 0.00 3/18/2014 2.05 0.83 2.83 0.00 3/19/2014 1.98 0.81 2.73 0.00 3/20/2014 1.90 0.77 2.63 0.00 3/21/2014 1.90 0.73 2.61 0.00 3/22/2014 1.92 0.75 2.60 0.00 3/23/2014 1.85 0.70 2.50 0.00 3/24/2014 1.90 0.66 2.56 0.00 3/25/2014 1.90 0.65 2.58 0.00 3/26/2014 1.88 0.63 2.54 0.00 3/27/2014 1.90 0.68 2.59 0.00 3/28/2014 1.95 0.72 2.60 0.00 3/29/2014 1.94 0.73 2.58 0.00 3/30/2014 1.91 0.71 2.56 0.00 3/31/2014 1.85 0.69 2.50 0.00 4/1/2014 1.82 0.65 2.45 0.00 4/2/2014 1.81 0.61 2.43 0.00 4/3/2014 1.83 0.63 2.45 0.00 4/4/2014 1.88 0.65 2.50 0.00 4/5/2014 1.88 0.68 2.51 0.00 4/6/2014 1.90 0.70 2.51 0.00 4/7/2014 1.97 0.70 2.58 0.00 4/8/2014 2.00 0.75 2.61 0.00 4/9/2014 1.97 0.75 2.60 0.00 4/10/2014 1.93 0.74 2.55 0.00 4/11/2014 1.90 0.70 2.52 0.00 4/12/2014 1.90 0.70 2.50 0.00 4/13/2014 1.93 0.67 2.53 0.00 4/14/2014 1.95 0.68 2.60 0.00 4/15/2014 1.96 0.70 2.62 0.00 4/16/2014 1.95 0.70 2.60 0.00 4/17/2014 1.95 0.70 2.58 0.00 4/18/2014 1.94 0.70 2.57 0.00 4/19/2014 1.99 0.74 2.61 0.00 4/20/2014 1.98 0.74 2.60 0.00 4/21/2014 1.93 0.71 2.53 0.00

4/22/2014 1.85 0.66 2.45 0.00 4/23/2014 1.80 0.58 2.38 0.00 4/24/2014 1.79 0.56 2.38 0.00 4/25/2014 1.84 0.59 2.41 0.00 4/26/2014 1.87 0.61 2.44 0.00 4/27/2014 1.90 0.65 2.49 0.00 4/28/2014 1.88 0.66 2.46 0.00 4/29/2014 1.88 0.67 2.43 0.00 4/30/2014 1.86 0.66 2.40 0.00 5/1/2014 1.86 0.65 2.40 0.00 5/2/2014 1.84 0.65 2.38 0.00 5/3/2014 1.88 0.64 2.38 0.00 5/4/2014 1.85 0.64 2.35 0.00 5/5/2014 1.84 0.63 2.35 0.00 5/6/2014 1.85 0.63 2.36 0.00 5/7/2014 1.84 0.63 2.34 0.00 5/8/2014 1.79 0.57 2.29 0.00 5/9/2014 1.77 0.54 2.28 0.00 5/10/2014 1.78 0.54 2.29 0.00 5/11/2014 1.78 0.54 2.31 0.00 5/12/2014 1.82 0.56 2.35 0.00 5/13/2014 1.84 0.60 2.35 0.00 5/14/2014 1.84 0.62 2.35 0.00 5/15/2014 1.81 0.59 2.32 0.00 5/16/2014 1.78 0.57 2.27 0.00 5/17/2014 1.73 0.55 2.22 0.00 5/18/2014 1.71 0.52 2.20 0.00 5/19/2014 1.70 0.48 2.18 0.00 5/20/2014 1.72 0.47 2.20 0.00 5/21/2014 1.70 0.45 2.21 0.00 5/22/2014 1.72 0.47 2.24 0.00 5/23/2014 1.78 0.51 2.27 0.00 5/24/2014 1.80 0.54 2.28 0.00 5/25/2014 1.82 0.53 2.34 0.00 5/26/2014 1.84 0.55 2.36 0.00 5/27/2014 1.80 0.55 2.32 0.00 5/28/2014 1.80 0.54 2.29 0.00 5/29/2014 1.78 0.54 2.27 0.00 5/30/2014 1.78 0.53 2.26 0.00 5/31/2014 1.79 0.54 2.27 0.00

6/1/2014 1.75 0.53 2.23 0.00 6/2/2014 1.71 0.48 2.21 0.00 6/3/2014 1.69 0.46 2.18 0.00 6/4/2014 1.65 0.44 2.14 0.00 6/5/2014 1.66 0.42 2.16 0.00 6/6/2014 1.66 0.42 2.16 0.00 6/7/2014 1.72 0.44 2.21 0.00 6/8/2014 1.68 0.45 2.18 0.00 6/9/2014 1.61 0.43 2.10 0.00 6/10/2014 1.66 0.43 2.14 0.00 6/11/2014 1.79 0.47 2.23 0.00 6/12/2014 1.77 0.51 2.23 0.00 6/13/2014 1.75 0.50 2.22 0.00 6/14/2014 1.67 0.46 2.15 0.00 6/15/2014 1.67 0.43 2.14 0.00 6/16/2014 1.70 0.43 2.18 0.00 6/17/2014 1.71 0.43 2.21 0.00 6/18/2014 1.70 0.43 2.19 0.00 6/19/2014 1.73 0.43 2.23 0.00 6/20/2014 1.78 0.49 2.28 0.00 6/21/2014 1.76 0.50 2.26 0.00 6/22/2014 1.77 0.50 2.27 0.00 6/23/2014 1.78 0.50 2.27 0.00 6/24/2014 1.78 0.49 2.25 0.00 6/25/2014 1.69 0.47 2.17 0.00 6/26/2014 1.75 0.43 2.18 0.00 6/27/2014 1.73 0.42 2.21 0.00 6/28/2014 1.76 0.43 2.25 0.00 6/29/2014 1.79 0.45 2.30 0.00 6/30/2014 1.80 0.50 2.32 0.00 7/1/2014 1.77 0.47 2.27 0.00 7/2/2014 1.76 0.45 2.27 0.00 7/3/2014 1.76 0.44 2.30 0.00 7/4/2014 1.73 0.43 2.27 0.00 7/5/2014 1.80 0.42 2.30 0.66 7/6/2014 1.82 0.38 2.30 0.00 7/7/2014 1.86 0.38 2.38 0.00 7/8/2014 1.90 0.43 2.54 0.00 7/9/2014 1.88 0.46 2.61 0.00 7/10/2014 1.84 0.51 2.59 0.00

7/11/2014 1.78 0.48 2.49 0.00 7/12/2014 1.76 0.45 2.42 0.00 7/13/2014 1.78 0.45 2.45 0.00 7/14/2014 1.83 0.51 2.52 0.00 7/15/2014 1.83 0.52 2.48 0.00 7/16/2014 1.83 0.53 2.45 0.00 7/17/2014 1.74 0.49 2.33 0.00 7/18/2014 1.69 0.45 2.27 0.00 7/19/2014 1.68 0.42 2.26 0.00 7/20/2014 1.72 0.43 2.28 0.00 7/21/2014 1.77 0.45 2.34 0.00 7/22/2014 1.76 0.46 2.33 0.00 7/23/2014 1.79 0.49 2.34 0.00 7/24/2014 1.72 0.48 2.27 0.00 7/25/2014 1.68 0.41 2.21 0.00 7/26/2014 1.68 0.38 2.21 0.00 7/27/2014 1.75 0.43 2.29 0.00 7/28/2014 1.80 0.45 2.33 0.55 0.30 0.00 7/29/2014 1.80 0.52 2.33 0.55 0.30 0.00 7/30/2014 1.80 0.53 2.32 0.55 0.31 0.00 7/31/2014 1.79 0.51 2.30 0.55 0.32 0.00 8/1/2014 1.81 0.50 2.31 0.57 0.33 0.00 8/2/2014 1.77 0.48 2.29 0.56 0.33 0.00 8/3/2014 1.76 0.46 2.28 0.57 0.33 0.00 8/4/2014 1.74 0.46 2.25 0.56 0.32 0.00 8/5/2014 1.69 0.44 2.21 0.55 0.28 0.00 8/6/2014 1.67 0.40 2.18 0.54 0.27 0.00 8/7/2014 1.66 0.39 2.19 0.53 0.27 0.00 8/8/2014 1.73 0.42 2.23 0.53 0.29 0.00 8/9/2014 1.73 0.43 2.26 0.53 0.30 0.00 8/10/2014 1.69 0.43 2.23 0.54 0.28 0.00 8/11/2014 1.72 0.43 2.25 0.54 0.30 0.00 8/12/2014 1.71 0.43 2.20 0.53 0.30 0.00 8/13/2014 1.69 0.38 2.19 0.53 0.28 0.00 8/14/2014 1.71 0.39 2.25 0.54 0.28 0.00 8/15/2014 1.76 0.43 2.29 0.54 0.33 0.00 8/16/2014 1.84 0.48 2.38 0.58 0.34 0.00 8/17/2014 1.80 0.48 2.35 0.60 0.35 0.00 8/18/2014 1.77 0.47 2.29 0.58 0.34 0.00 8/19/2014 1.80 0.45 2.33 0.56 0.34 0.00

8/20/2014 1.82 0.46 2.37 0.57 0.34 0.00 8/21/2014 1.77 0.45 2.35 0.55 0.33 0.00 8/22/2014 1.78 0.44 2.35 0.56 0.34 0.00 8/23/2014 1.76 0.45 2.34 0.58 0.34 0.00 8/24/2014 1.76 0.47 2.34 0.59 0.34 0.00 8/25/2014 1.76 0.46 2.29 0.57 0.34 0.00 8/26/2014 1.79 0.45 2.32 0.58 0.34 0.00 8/27/2014 1.74 0.45 2.31 0.58 0.34 0.00 8/28/2014 1.75 0.44 2.29 0.57 0.33 0.00 8/29/2014 1.76 0.44 2.31 0.56 0.33 0.00 8/30/2014 1.84 0.50 2.38 0.60 0.35 0.00 8/31/2014 1.80 0.52 2.35 0.62 0.35 0.00 9/1/2014 1.73 0.50 2.26 0.62 0.35 0.00 9/2/2014 1.74 0.46 2.23 0.60 0.35 0.00 9/3/2014 1.79 0.45 2.28 0.59 0.34 0.00 9/4/2014 1.86 0.50 2.38 0.63 0.38 0.00 9/5/2014 1.83 0.53 2.35 0.64 0.39 0.00 9/6/2014 1.85 0.54 2.36 0.65 0.40 0.00 9/7/2014 1.85 0.55 2.35 0.66 0.42 0.00 9/8/2014 1.81 0.55 2.29 0.68 0.42 0.00 9/9/2014 1.79 0.55 2.26 0.68 0.41 0.00 9/10/2014 1.76 0.52 2.22 0.66 0.40 0.00 9/11/2014 1.76 0.48 2.21 0.66 0.39 0.00 9/12/2014 1.75 0.47 2.20 0.65 0.38 0.00 9/13/2014 1.80 0.48 2.23 0.65 0.40 0.00 9/14/2014 1.83 0.50 2.26 0.65 0.43 0.00 9/15/2014 1.87 0.50 2.30 0.68 0.42 0.00 9/16/2014 1.86 0.51 2.34 0.66 0.43 0.00 9/17/2014 1.83 0.48 2.29 0.61 0.40 0.00 9/18/2014 1.85 0.45 2.31 0.62 0.40 0.00 9/19/2014 1.84 0.46 2.34 0.62 0.41 0.00 9/20/2014 1.87 0.50 2.40 0.64 0.45 0.00 9/21/2014 1.83 0.55 2.35 0.66 0.46 0.00 9/22/2014 1.82 0.54 2.31 0.65 0.44 0.00 9/23/2014 1.84 0.50 2.31 0.63 0.42 0.98 9/24/2014 1.98 0.50 2.48 0.63 0.45 0.00 9/25/2014 1.89 0.53 2.55 0.65 0.45 0.00 9/26/2014 1.85 0.53 2.50 0.65 0.44 0.00 9/27/2014 1.85 0.50 2.48 0.64 0.45 0.00 9/28/2014 1.81 0.47 2.41 0.63 0.43 0.00

9/29/2014 1.84 0.48 2.41 0.63 0.44 0.00 9/30/2014 1.81 0.48 2.41 0.63 0.43 0.00 10/1/2014 1.84 0.50 2.43 0.64 0.46 0.00 10/2/2014 1.83 0.53 2.39 0.65 0.48 0.00 10/3/2014 1.85 0.53 2.39 0.65 0.48 0.00 10/4/2014 1.86 0.54 2.39 0.66 0.48 0.00 10/5/2014 1.87 0.54 2.39 0.66 0.48 0.00 10/6/2014 1.86 0.54 2.40 0.66 0.48 0.00 10/7/2014 1.87 0.54 2.43 0.66 0.49 0.00 10/8/2014 1.88 0.54 2.42 0.68 0.48 0.00 10/9/2014 1.86 0.55 2.42 0.67 0.48 0.00 10/10/2014 1.80 0.51 2.35 0.66 0.47 0.00 10/11/2014 1.86 0.52 2.39 0.65 0.48 0.00 10/12/2014 1.86 0.55 2.42 0.67 0.50 0.00 10/13/2014 1.89 0.58 2.42 0.70 0.52 0.00 10/14/2014 1.85 0.56 2.36 0.70 0.52 0.00 10/15/2014 1.88 0.55 2.38 0.70 0.51 0.00 10/16/2014 1.85 0.55 2.36 0.69 0.51 0.00 10/17/2014 1.85 0.54 2.35 0.68 0.51 0.00 10/18/2014 1.86 0.55 2.35 0.70 0.52 0.00 10/19/2014 1.88 0.54 2.35 0.68 0.52 0.00 10/20/2014 1.88 0.54 2.38 0.69 0.51 0.00 10/21/2014 1.88 0.55 2.40 0.69 0.53 0.00 10/22/2014 1.87 0.55 2.38 0.70 0.53 0.00 10/23/2014 1.88 0.55 2.39 0.70 0.56 0.00 10/24/2014 1.85 0.54 2.35 0.69 0.55 0.00 10/25/2014 1.85 0.54 2.35 0.70 0.55 0.00 10/26/2014 1.86 0.54 2.36 0.70 0.55 0.00 10/27/2014 1.88 0.55 2.38 0.70 0.55 0.00 10/28/2014 1.90 0.55 2.42 0.70 0.57 0.00 10/29/2014 1.89 0.56 2.42 0.70 0.57 0.00 10/30/2014 1.93 0.58 2.46 0.72 0.59 0.00 10/31/2014 1.90 0.60 2.43 0.74 0.60 0.00 11/1/2014 1.86 0.56 2.37 0.71 0.57 0.00 11/2/2014 1.84 0.54 2.34 0.70 0.55 0.00 11/3/2014 1.85 0.53 2.37 0.69 0.55 0.00 11/4/2014 1.87 0.54 2.38 0.72 0.58 0.00 11/5/2014 1.93 0.59 2.44 0.73 0.60 0.00 11/6/2014 1.91 0.62 2.40 0.76 0.64 0.00 11/7/2014 1.90 0.60 2.37 0.76 0.63 0.00

11/8/2014 1.92 0.60 2.38 0.75 0.63 0.00 11/9/2014 1.90 0.59 2.38 0.75 0.63 0.00 11/10/2014 1.90 0.59 2.38 0.75 0.63 0.00 11/11/2014 1.90 0.59 2.37 0.75 0.64 0.00 11/12/2014 1.88 0.58 2.30 0.75 0.64 0.00 11/13/2014 1.86 0.57 2.26 0.75 0.61 0.00 11/14/2014 1.84 0.55 2.24 0.75 0.61 0.00 11/15/2014 1.88 0.56 2.26 0.74 0.62 0.00 11/16/2014 1.91 0.59 2.30 0.75 0.64 0.00 11/17/2014 1.92 0.61 2.32 0.77 0.65 0.00 11/18/2014 1.93 0.66 2.32 0.79 0.65 0.00 11/19/2014 1.87 0.65 2.24 0.81 0.66 0.00 11/20/2014 1.85 0.57 2.18 0.77 0.65 0.00 11/21/2014 1.84 0.55 2.17 0.76 0.64 0.00 11/22/2014 1.82 0.54 2.15 0.73 0.61 0.00 11/23/2014 1.85 0.53 2.16 0.73 0.61 0.00 11/24/2014 1.86 0.55 2.17 0.75 0.63 0.00 11/25/2014 1.85 0.55 2.15 0.75 0.63 0.00 11/26/2014 1.89 0.55 2.18 0.75 0.63 0.00 11/27/2014 1.86 0.55 2.18 0.75 0.63 0.00 11/28/2014 1.87 0.56 2.19 0.75 0.63 0.00 11/29/2014 1.88 0.55 2.18 0.75 0.64 0.00 11/30/2014 1.87 0.56 2.19 0.75 0.64 0.00 12/1/2014 1.88 0.57 2.20 0.76 0.64 0.00 12/2/2014 1.90 0.58 2.20 0.76 0.65 0.00 12/3/2014 1.90 0.59 2.21 0.77 0.65 0.00 12/4/2014 1.91 0.60 2.23 0.78 0.65 0.00 12/5/2014 1.92 0.60 2.24 0.78 0.65 0.00 12/6/2014 1.90 0.62 2.20 0.80 0.66 0.00 12/7/2014 1.90 0.62 2.19 0.80 0.66 0.00 12/8/2014 1.88 0.60 2.17 0.78 0.65 0.00 12/9/2014 1.86 0.58 2.16 0.78 0.65 0.00 12/10/2014 1.89 0.58 2.17 0.77 0.65 0.00 12/11/2014 1.90 0.58 2.19 0.78 0.65 0.00 12/12/2014 1.90 0.58 2.19 0.77 0.65 0.00 12/13/2014 1.88 0.57 2.17 0.77 0.65 0.00 12/14/2014 1.90 0.57 2.20 0.75 0.65 0.00 12/15/2014 1.88 0.55 2.18 0.75 0.64 0.00 12/16/2014 1.87 0.55 2.18 0.75 0.64 0.00 12/17/2014 1.87 0.55 2.18 0.74 0.64 0.00

12/18/2014 1.91 0.55 2.21 0.75 0.64 0.00 12/19/2014 1.90 0.55 2.22 0.75 0.65 0.00 12/20/2014 1.91 0.58 2.23 0.75 0.65 0.00 12/21/2014 1.93 0.59 2.25 0.76 0.65 0.00 12/22/2014 1.90 0.59 2.23 0.76 0.65 0.00 12/23/2014 1.90 0.58 2.23 0.76 0.65 0.00 12/24/2014 1.90 0.58 2.23 0.77 0.65 0.00 12/25/2014 1.90 0.58 2.23 0.77 0.66 0.00 12/26/2014 1.91 0.57 2.23 0.77 0.65 0.00 12/27/2014 1.94 0.59 2.25 0.78 0.65 0.00 12/28/2014 1.89 0.57 2.22 0.76 0.65 0.00 12/29/2014 1.93 0.58 2.25 0.76 0.65 0.00 12/30/2014 1.93 0.58 2.24 0.77 0.65 0.00 12/31/2014 1.96 0.60 2.26 0.77 0.66 0.00 1/1/2015 1.95 0.56 2.24 0.76 0.65 0.00 1/2/2015 1.95 0.57 2.27 0.75 0.64 0.00 1/3/2015 1.96 0.60 2.30 0.76 0.65 0.00 1/4/2015 1.96 0.61 2.29 0.77 0.65 0.00 1/5/2015 1.98 0.63 2.30 0.78 0.66 0.00 1/6/2015 1.98 0.63 2.28 0.73 0.65 0.00 1/7/2015 2.00 0.62 2.32 0.76 0.66 0.00 1/8/2015 2.01 0.64 2.35 0.80 0.68 0.00 1/9/2015 2.02 0.65 2.37 0.79 0.68 0.00 1/10/2015 2.04 0.67 2.38 0.79 0.70 0.00 1/11/2015 2.05 0.67 2.38 0.80 0.73 0.00 1/12/2015 1.99 0.66 2.35 0.77 0.71 0.00 1/13/2015 1.97 0.64 2.30 0.75 0.68 0.00 1/14/2015 2.00 0.65 2.35 0.77 0.68 0.00 1/15/2015 2.01 0.65 2.37 0.77 0.70 0.00 1/16/2015 2.00 0.66 2.37 0.75 0.68 0.00 1/17/2015 2.00 0.66 2.35 0.75 0.67 0.00 1/18/2015 1.97 0.63 2.30 0.70 0.66 0.00 1/19/2015 1.94 0.60 2.27 0.68 0.66 0.00 1/20/2015 1.99 0.60 2.33 0.73 0.65 0.00 1/21/2015 2.01 0.65 2.37 0.75 0.67 0.00 1/22/2015 2.08 0.70 2.43 0.81 0.73 0.00 1/23/2015 2.03 0.70 2.38 0.76 0.70 0.00 1/24/2015 2.08 0.70 2.43 0.79 0.73 0.00 1/25/2015 3.31 0.70 2.63 0.71 0.70 4.59 1/26/2015 6.65 0.74 3.66 0.76 0.68 7.54

1/27/2015 5.70 0.86 9.11 0.82 0.72 0.00 1/28/2015 4.94 1.03 9.17 0.79 0.73 0.00 1/29/2015 4.64 1.12 8.81 0.77 0.72 0.00 1/30/2015 4.42 1.20 8.92 0.75 0.70 0.00 1/31/2015 4.23 1.26 8.77 0.77 0.70 0.00 2/1/2015 3.92 1.33 8.61 0.75 0.69 0.00 2/2/2015 3.53 1.33 8.21 0.74 0.67 0.00 2/3/2015 3.17 1.35 8.00 0.74 0.67 0.00 2/4/2015 2.88 1.39 7.79 0.73 0.67 0.00 2/5/2015 2.73 1.43 7.63 0.72 0.67 0.00 2/6/2015 2.60 1.43 7.39 0.78 0.66 0.00 2/7/2015 2.54 1.44 7.21 0.78 0.67 0.00 2/8/2015 2.48 1.45 7.05 0.79 0.67 0.00 2/9/2015 2.43 1.42 6.89 0.78 0.67 0.00 2/10/2015 2.43 1.40 6.72 0.79 0.67 0.00 2/11/2015 2.37 1.35 6.54 0.79 0.67 0.00 2/12/2015 2.37 1.30 6.39 0.80 0.67 0.00 2/13/2015 2.32 1.23 6.12 0.79 0.67 0.00 2/14/2015 2.27 1.18 5.85 0.75 0.66 0.00 2/15/2015 2.26 1.11 5.70 0.75 0.66 0.00 2/16/2015 2.27 1.08 5.61 0.75 0.65 0.00 2/17/2015 2.26 1.06 5.47 0.75 0.65 0.00 2/18/2015 2.26 1.05 5.44 0.78 0.65 0.00 2/19/2015 2.28 1.07 5.40 0.80 0.67 0.00 2/20/2015 2.26 1.06 5.19 0.79 0.68 0.00 2/21/2015 2.23 1.04 5.00 0.81 0.67 0.00 2/22/2015 2.24 1.03 4.87 0.81 0.68 0.00 2/23/2015 2.22 1.02 4.69 0.82 0.68 0.00 2/24/2015 2.20 1.00 4.50 0.81 0.68 0.00 2/25/2015 2.12 0.93 4.25 0.77 0.67 0.00 2/26/2015 2.13 0.89 4.17 0.75 0.66 0.00 2/27/2015 2.11 0.86 4.08 0.77 0.65 0.00 2/28/2015 2.10 0.82 4.01 0.74 0.65 0.00 3/1/2015 2.15 0.82 4.01 0.76 0.65 0.00 3/2/2015 2.14 0.82 4.00 0.76 0.65 0.00 3/3/2015 2.17 0.83 3.98 0.79 0.65 0.00 3/4/2015 2.18 0.84 3.96 0.80 0.66 0.00 3/5/2015 2.19 0.88 3.86 0.83 0.67 0.00 3/6/2015 2.19 0.90 3.74 0.86 0.70 0.00 3/7/2015 2.18 0.88 3.60 0.84 0.70 0.00

3/8/2015 2.16 0.85 3.52 0.84 0.70 0.00 3/9/2015 2.18 0.85 3.45 0.85 0.70 0.00 3/10/2015 2.16 0.82 3.40 0.85 0.71 0.00 3/11/2015 2.15 0.81 3.32 0.85 0.70 0.00 3/12/2015 2.10 0.79 3.22 0.82 0.69 0.00 3/13/2015 2.11 0.79 3.18 0.85 0.69 0.00 3/14/2015 2.10 0.79 3.10 0.87 0.70 0.00 3/15/2015 2.08 0.79 3.00 0.84 0.70 0.00 3/16/2015 2.13 0.79 3.00 0.92 0.72 0.00 3/17/2015 2.10 0.82 2.90 0.90 0.73 0.00 3/18/2015 2.03 0.74 2.76 0.83 0.70 0.00 3/19/2015 2.07 0.73 2.82 0.87 0.68 0.00 3/20/2015 2.06 0.75 2.75 0.86 0.70 0.00 3/21/2015 2.10 0.75 2.76 0.86 0.70 0.00 3/22/2015 2.10 0.76 2.78 0.88 0.70 0.00 3/23/2015 2.06 0.72 2.70 0.85 0.70 0.00 3/24/2015 2.00 0.67 2.64 0.80 0.67 0.00 3/25/2015 2.00 0.65 2.62 0.82 0.67 0.00 3/26/2015 2.00 0.66 2.63 0.82 0.66 0.00 3/27/2015 2.02 0.67 2.60 0.83 0.67 0.00 3/28/2015 2.02 0.66 2.58 0.85 0.67 0.66 3/29/2015 2.28 0.66 2.71 0.82 0.67 1.31 3/30/2015 2.52 0.67 3.08 0.81 0.65 0.00 3/31/2015 2.43 0.70 3.37 0.82 0.65 0.00 4/1/2015 2.31 0.73 3.55 0.83 0.65 0.00 4/2/2015 2.27 0.76 3.60 0.82 0.65 0.00 4/3/2015 2.24 0.76 3.58 0.83 0.66 0.00 4/4/2015 2.19 0.74 3.45 0.80 0.66 0.00 4/5/2015 2.17 0.70 3.37 0.80 0.65 0.00 4/6/2015 2.17 0.69 3.31 0.79 0.64 0.00 4/7/2015 2.18 0.70 3.30 0.81 0.64 0.00 4/8/2015 2.25 0.76 3.33 0.87 0.66 0.00 4/9/2015 2.22 0.79 3.23 0.88 0.69 0.00 4/10/2015 2.22 0.80 3.16 0.87 0.70 0.00 4/11/2015 2.18 0.76 3.09 0.85 0.69 0.00 4/12/2015 2.19 0.75 3.05 0.85 0.68 0.00 4/13/2015 2.20 0.78 3.04 0.87 0.70 0.00 4/14/2015 2.15 0.76 2.95 0.85 0.70 0.00 4/15/2015 2.08 0.70 2.85 0.82 0.68 0.00 4/16/2015 2.03 0.65 2.75 0.77 0.65 0.00

4/17/2015 1.98 0.61 2.71 0.75 0.61 0.00 4/18/2015 2.00 0.59 2.72 0.76 0.60 0.00 4/19/2015 2.00 0.61 2.70 0.76 0.60 0.00 4/20/2015 2.00 0.61 2.67 0.77 0.60 0.00 4/21/2015 2.00 0.61 2.64 0.78 0.60 0.00 4/22/2015 2.02 0.63 2.65 0.80 0.62 0.00 4/23/2015 2.06 0.65 2.67 0.84 0.63 0.00 4/24/2015 2.06 0.68 2.63 0.84 0.65 0.00 4/25/2015 2.05 0.69 2.61 0.85 0.66 0.00 4/26/2015 1.99 0.66 2.51 0.82 0.65 0.00 4/27/2015 1.95 0.62 2.43 0.79 0.63 0.00 4/28/2015 1.93 0.58 2.38 0.77 0.61 0.00 4/29/2015 1.92 0.56 2.36 0.74 0.60 0.00 4/30/2015 1.90 0.54 2.35 0.75 0.60 0.00 5/1/2015 1.99 0.57 2.42 0.80 0.60 0.00 5/2/2015 2.03 0.62 2.48 0.83 0.62 0.00 5/3/2015 2.05 0.65 2.51 0.85 0.65 0.00 5/4/2015 2.02 0.66 2.46 0.84 0.65 0.00 5/5/2015 2.01 0.66 2.41 0.85 0.65 0.00 5/6/2015 1.98 0.64 2.37 0.83 0.65 0.00 5/7/2015 1.97 0.64 2.33 0.82 0.65 0.00 5/8/2015 1.96 0.62 2.30 0.82 0.65 0.00 5/9/2015 1.95 0.60 2.28 0.82 0.65 0.00 5/10/2015 1.93 0.57 2.25 0.79 0.64 0.00 5/11/2015 1.89 0.55 2.21 0.78 0.60 0.00 5/12/2015 1.89 0.54 2.20 0.78 0.60 0.00 5/13/2015 1.88 0.54 2.18 0.76 0.60 0.00 5/14/2015 1.87 0.51 2.17 0.75 0.60 0.00 5/15/2015 1.90 0.53 2.19 0.78 0.60 0.00 5/16/2015 1.94 0.55 2.24 0.81 0.60 0.00 5/17/2015 1.96 0.58 2.24 0.83 0.62 0.00 5/18/2015 1.94 0.57 2.20 0.81 0.62 0.00 5/19/2015 1.93 0.55 2.20 0.80 0.60 0.00 5/20/2015 1.94 0.55 2.20 0.80 0.60 0.00 5/21/2015 1.95 0.55 2.21 0.80 0.61 0.00 5/22/2015 1.95 0.56 2.22 0.81 0.61 0.00 5/23/2015 1.93 0.55 2.18 0.79 0.60 0.00 5/24/2015 1.94 0.55 2.20 0.79 0.60 0.00 5/25/2015 1.87 0.53 2.12 0.75 0.60 0.00 5/26/2015 1.88 0.48 2.13 0.74 0.60 0.00

5/27/2015 1.90 0.48 2.16 0.74 0.60 0.00 5/28/2015 1.93 0.51 2.18 0.78 0.60 0.00 5/29/2015 1.93 0.53 2.20 0.78 0.60 0.00 5/30/2015 1.92 0.54 2.18 0.78 0.60 0.00 5/31/2015 1.84 0.50 2.10 0.73 0.60 0.00 6/1/2015 1.80 0.45 2.05 0.68 0.58 0.00 6/2/2015 1.83 0.43 2.07 0.67 0.53 0.00 6/3/2015 1.82 0.41 2.08 0.67 0.53 0.00 6/4/2015 1.87 0.43 2.13 0.69 0.53 0.00 6/5/2015 1.93 0.45 2.21 0.73 0.55 0.00 6/6/2015 1.93 0.48 2.21 0.73 0.56 0.00 6/7/2015 1.92 0.49 2.19 0.73 0.57 0.00 6/8/2015 1.83 0.47 2.10 0.70 0.57 0.00 6/9/2015 1.83 0.41 2.10 0.68 0.52 0.00 6/10/2015 1.81 0.40 2.10 0.67 0.50 0.00 6/11/2015 1.88 0.43 2.15 0.68 0.52 0.00 6/12/2015 1.92 0.44 2.22 0.72 0.53 0.00 6/13/2015 1.98 0.53 2.28 0.78 0.58 0.00 6/14/2015 1.95 0.53 2.23 0.79 0.60 0.00 6/15/2015 1.90 0.53 2.18 0.74 0.60 0.00 6/16/2015 1.83 0.48 2.11 0.71 0.58 0.00 6/17/2015 1.85 0.45 2.11 0.71 0.56 0.00 6/18/2015 1.86 0.46 2.12 0.71 0.57 0.00 6/19/2015 1.88 0.47 2.14 0.74 0.57 0.00 6/20/2015 1.88 0.49 2.13 0.75 0.58 0.00 6/21/2015 1.88 0.50 2.13 0.75 0.58 0.00 6/22/2015 1.87 0.49 2.12 0.75 0.60 0.00 6/23/2015 1.84 0.48 2.09 0.73 0.60 0.00 6/24/2015 1.80 0.45 2.03 0.68 0.57 0.00 6/25/2015 1.76 0.39 1.99 0.68 0.51 0.00 6/26/2015 1.78 0.38 2.02 0.67 0.50 0.00 6/27/2015 1.82 0.37 2.05 0.68 0.50 0.00 6/28/2015 1.82 0.40 2.07 0.69 0.52 0.00 6/29/2015 1.80 0.41 2.07 0.69 0.52 0.00 6/30/2015 1.81 0.43 2.07 0.69 0.52 0.00 7/1/2015 1.82 0.44 2.06 0.70 0.53 0.00 7/2/2015 1.83 0.44 2.08 0.72 0.54 0.00 7/3/2015 1.88 0.46 2.13 0.77 0.55 0.00 7/4/2015 1.90 0.48 2.14 0.77 0.60 0.00 7/5/2015 1.91 0.50 2.14 0.78 0.60 0.00

7/6/2015 1.91 0.51 2.14 0.78 0.60 0.00 7/7/2015 1.93 0.53 2.18 0.80 0.60 0.00 7/8/2015 1.93 0.53 2.17 0.80 0.60 0.00 7/9/2015 1.93 0.53 2.16 0.80 0.60 0.00 7/10/2015 1.93 0.54 2.15 0.80 0.60 0.00 7/11/2015 1.85 0.50 2.09 0.75 0.60 0.00 7/12/2015 1.84 0.45 2.06 0.73 0.59 0.00 7/13/2015 1.85 0.42 2.05 0.70 0.55 0.00 7/14/2015 1.87 0.40 2.06 0.70 0.54 0.00 7/15/2015 1.90 0.37 2.08 0.68 0.50 0.00 7/16/2015 1.89 0.37 2.10 0.67 0.49 0.00 7/17/2015 1.91 0.39 2.16 0.68 0.49 0.00 7/18/2015 1.91 0.42 2.18 0.71 0.52 0.00 7/19/2015 1.91 0.44 2.18 0.71 0.53 0.00 7/20/2015 1.85 0.43 2.12 0.69 0.53 0.00 7/21/2015 1.90 0.42 2.16 0.70 0.50 0.00 7/22/2015 1.95 0.40 2.18 0.71 0.50 0.00 7/23/2015 1.98 0.43 2.24 0.73 0.52 0.00 7/24/2015 1.98 0.43 2.29 0.73 0.53 0.00 7/25/2015 1.94 0.43 2.27 0.69 0.51 0.00 7/26/2015 1.99 0.45 2.35 0.75 0.53 0.00 7/27/2015 1.93 0.47 2.28 0.71 0.54 0.00 7/28/2015 1.78 0.45 2.13 0.66 0.52 0.00

Appendix B: Soil moisture dynamics and rainfall patterns at Kleinberg.

Gravel plain Kleinberg Gravel plain Kleinberg Time Bare soil Rain Time Bare soil Rain Depth 5 cm Depth 5 cm Unit m^3/m^3 mm Unit m^3/m^3 mm 1/2/2014 0.51 0.00 10/19/2014 0.43 0.00 1/3/2014 0.42 0.00 10/20/2014 0.43 0.00 1/4/2014 0.46 0.00 10/21/2014 0.42 0.00 1/5/2014 0.51 0.00 10/22/2014 0.43 0.00 1/6/2014 0.41 0.00 10/23/2014 0.45 0.00 1/7/2014 0.38 0.00 10/24/2014 0.43 0.00 1/8/2014 0.44 0.00 10/25/2014 0.43 0.00 1/9/2014 0.40 0.00 10/26/2014 0.44 0.00 1/10/2014 0.41 0.00 10/27/2014 0.46 0.00 1/11/2014 0.45 0.00 10/28/2014 0.45 0.00 1/12/2014 0.43 0.00 10/29/2014 0.47 0.00 1/13/2014 0.44 0.00 10/30/2014 0.46 0.00 1/14/2014 0.43 0.00 10/31/2014 0.43 0.00 1/15/2014 0.46 0.00 11/1/2014 0.41 0.00 1/16/2014 0.66 3.51 11/2/2014 0.45 0.00 1/17/2014 0.76 0.00 11/3/2014 0.43 0.00 1/18/2014 0.63 0.00 11/4/2014 0.41 0.00 1/19/2014 0.55 0.00 11/5/2014 0.48 0.00 1/20/2014 0.53 0.00 11/6/2014 0.49 0.00 1/21/2014 0.46 0.00 11/7/2014 0.55 0.00 1/22/2014 0.48 0.00 11/8/2014 0.63 0.00 1/23/2014 0.48 0.00 11/9/2014 0.59 0.00 1/24/2014 0.47 0.00 11/10/2014 0.56 0.00 1/25/2014 0.43 0.00 11/11/2014 0.53 0.00 1/26/2014 0.44 0.00 11/12/2014 0.53 0.00 1/27/2014 0.40 0.00 11/13/2014 0.52 0.00 1/28/2014 0.45 0.00 11/14/2014 0.50 0.00 1/29/2014 0.49 0.00 11/15/2014 0.54 0.00 1/30/2014 0.47 0.00 11/16/2014 0.60 0.00 1/31/2014 0.45 0.00 11/17/2014 0.64 0.00 2/1/2014 0.43 0.00 11/18/2014 0.65 0.00 2/2/2014 0.52 0.00 11/19/2014 0.59 0.00 2/3/2014 0.54 0.00 11/20/2014 0.69 0.00

2/4/2014 0.51 0.44 11/21/2014 0.69 0.00 2/5/2014 0.58 0.00 11/22/2014 0.68 0.00 2/6/2014 0.53 0.00 11/23/2014 0.66 0.00 2/7/2014 0.53 0.00 11/24/2014 0.62 0.00 2/8/2014 0.50 0.00 11/25/2014 0.68 0.00 2/9/2014 0.42 0.00 11/26/2014 0.70 0.00 2/10/2014 0.50 0.00 11/27/2014 0.68 0.00 2/11/2014 0.70 4.39 11/28/2014 0.65 0.00 2/12/2014 0.89 0.00 11/29/2014 0.71 0.00 2/13/2014 0.84 0.00 11/30/2014 0.67 0.00 2/14/2014 0.73 0.00 12/1/2014 0.67 0.00 2/15/2014 0.64 0.00 12/2/2014 0.68 0.00 2/16/2014 0.60 0.00 12/3/2014 0.69 0.00 2/17/2014 0.53 0.00 12/4/2014 0.68 0.00 2/18/2014 0.57 0.00 12/5/2014 0.65 0.00 2/19/2014 0.53 0.00 12/6/2014 0.65 0.00 2/20/2014 0.55 0.00 12/7/2014 0.68 0.00 2/21/2014 0.48 0.00 12/8/2014 0.64 0.00 2/22/2014 0.45 0.00 12/9/2014 0.68 0.00 2/23/2014 0.45 0.00 12/10/2014 0.70 0.00 2/24/2014 0.47 0.00 12/11/2014 0.70 0.00 2/25/2014 0.48 0.00 12/12/2014 0.69 0.44 2/26/2014 0.46 0.88 12/13/2014 0.67 0.00 2/27/2014 0.58 0.00 12/14/2014 0.70 0.00 2/28/2014 0.57 0.00 12/15/2014 0.65 0.00 3/1/2014 0.53 0.00 12/16/2014 0.61 0.00 3/2/2014 0.47 0.00 12/17/2014 0.70 0.00 3/3/2014 0.45 0.00 12/18/2014 0.70 0.00 3/4/2014 0.47 0.00 12/19/2014 0.65 0.00 3/5/2014 0.52 0.00 12/20/2014 0.68 0.00 3/6/2014 0.53 0.00 12/21/2014 0.70 0.00 3/7/2014 0.52 0.00 12/22/2014 0.67 0.00 3/8/2014 0.53 0.00 12/23/2014 0.68 0.00 3/9/2014 0.53 0.00 12/24/2014 0.67 0.00 3/10/2014 0.49 0.00 12/25/2014 0.70 0.00 3/11/2014 0.51 0.00 12/26/2014 0.70 0.00 3/12/2014 0.46 0.00 12/27/2014 0.73 0.00 3/13/2014 0.46 0.00 12/28/2014 0.67 0.00 3/14/2014 0.48 0.00 12/29/2014 0.69 0.00 3/15/2014 0.49 0.00 12/30/2014 0.68 0.00

3/16/2014 0.50 0.00 12/31/2014 0.74 0.00 3/17/2014 0.51 0.00 1/1/2015 0.73 0.00 3/18/2014 0.52 0.00 1/2/2015 0.75 0.00 3/19/2014 0.47 0.00 1/3/2015 0.75 0.00 3/20/2014 0.44 0.00 1/4/2015 0.74 0.00 3/21/2014 0.46 0.00 1/5/2015 0.75 0.00 3/22/2014 0.46 0.00 1/6/2015 0.72 0.44 3/23/2014 0.42 0.88 1/7/2015 2.81 10.54 3/24/2014 0.69 0.00 1/8/2015 3.38 0.00 3/25/2014 0.60 0.00 1/9/2015 2.88 0.00 3/26/2014 0.47 0.00 1/10/2015 2.27 0.00 3/27/2014 0.45 0.00 1/11/2015 1.60 0.00 3/28/2014 0.46 0.00 1/12/2015 1.31 0.00 3/29/2014 0.46 0.00 1/13/2015 1.19 0.00 3/30/2014 0.48 0.00 1/14/2015 1.22 0.00 3/31/2014 0.46 0.00 1/15/2015 1.15 0.00 4/1/2014 0.46 0.00 1/16/2015 1.08 0.00 4/2/2014 0.45 0.00 1/17/2015 1.02 0.00 4/3/2014 0.44 0.00 1/18/2015 0.99 0.00 4/4/2014 0.45 0.00 1/19/2015 0.95 0.00 4/5/2014 0.48 0.00 1/20/2015 1.01 0.00 4/6/2014 0.67 2.20 1/21/2015 0.99 0.00 4/7/2014 0.77 0.00 1/22/2015 0.97 0.00 4/8/2014 0.67 0.00 1/23/2015 0.92 0.00 4/9/2014 0.62 0.00 1/24/2015 1.02 0.00 4/10/2014 0.51 0.00 1/25/2015 4.11 13.18 4/11/2014 0.43 0.00 1/26/2015 4.35 0.00 4/12/2014 0.44 0.88 1/27/2015 3.83 0.00 4/13/2014 0.56 0.00 1/28/2015 3.24 0.00 4/14/2014 0.55 0.00 1/29/2015 2.69 0.00 4/15/2014 0.53 0.00 1/30/2015 2.08 0.00 4/16/2014 0.48 0.00 1/31/2015 1.70 0.00 4/17/2014 0.46 0.00 2/1/2015 1.40 0.00 4/18/2014 0.51 0.00 2/2/2015 1.27 0.00 4/19/2014 0.57 0.00 2/3/2015 1.23 0.00 4/20/2014 0.51 0.00 2/4/2015 1.13 0.00 4/21/2014 0.43 0.00 2/5/2015 1.10 0.44 4/22/2014 0.43 0.00 2/6/2015 1.09 0.00 4/23/2014 0.38 0.00 2/7/2015 1.05 0.00 4/24/2014 0.40 0.00 2/8/2015 1.02 0.00

4/25/2014 0.43 0.00 2/9/2015 0.98 0.00 4/26/2014 0.45 0.00 2/10/2015 0.98 0.00 4/27/2014 0.45 0.00 2/11/2015 0.96 0.00 4/28/2014 0.42 0.00 2/12/2015 0.94 0.00 4/29/2014 0.35 0.00 2/13/2015 0.90 0.00 4/30/2014 0.35 0.00 2/14/2015 0.90 0.00 5/1/2014 0.40 0.00 2/15/2015 0.91 0.00 5/2/2014 0.41 0.00 2/16/2015 0.93 0.00 5/3/2014 0.42 0.00 2/17/2015 0.93 0.00 5/4/2014 0.35 0.00 2/18/2015 0.90 0.00 5/5/2014 0.39 0.00 2/19/2015 0.89 0.00 5/6/2014 0.42 0.00 2/20/2015 0.86 0.00 5/7/2014 0.41 0.00 2/21/2015 0.88 0.00 5/8/2014 0.42 0.00 2/22/2015 0.84 0.00 5/9/2014 0.44 0.00 2/23/2015 0.84 0.00 5/10/2014 0.45 0.00 2/24/2015 0.83 0.00 5/11/2014 0.44 0.00 2/25/2015 0.84 0.00 5/12/2014 0.45 0.00 2/26/2015 0.82 0.00 5/13/2014 0.40 0.00 2/27/2015 0.83 0.00 5/14/2014 0.38 0.00 2/28/2015 0.82 0.00 5/15/2014 0.34 0.00 3/1/2015 0.83 0.00 5/16/2014 0.34 0.00 3/2/2015 0.83 0.00 5/17/2014 0.35 0.00 3/3/2015 0.85 0.00 5/18/2014 0.40 0.00 3/4/2015 0.87 0.00 5/19/2014 0.41 0.00 3/5/2015 0.85 0.00 5/20/2014 0.44 0.00 3/6/2015 0.86 0.00 5/21/2014 0.42 0.00 3/7/2015 0.84 0.00 5/22/2014 0.39 0.00 3/8/2015 0.84 0.00 5/23/2014 0.44 0.00 3/9/2015 0.82 0.00 5/24/2014 0.51 0.00 3/10/2015 0.84 0.00 5/25/2014 0.53 0.00 3/11/2015 0.81 0.00 5/26/2014 0.50 0.00 3/12/2015 0.83 0.00 5/27/2014 0.44 0.00 3/13/2015 0.80 0.00 5/28/2014 0.40 0.00 3/14/2015 0.78 0.00 5/29/2014 0.37 0.00 3/15/2015 0.79 0.00 5/30/2014 0.38 0.00 3/16/2015 0.76 0.00 5/31/2014 0.38 0.00 3/17/2015 0.78 0.00 6/1/2014 0.43 0.00 3/18/2015 0.78 0.00 6/2/2014 0.44 0.00 3/19/2015 0.78 0.00 6/3/2014 0.43 0.00 3/20/2015 0.75 0.00

6/4/2014 0.41 0.00 3/21/2015 0.80 0.00 6/5/2014 0.42 0.00 3/22/2015 0.78 0.00 6/6/2014 0.41 0.00 3/23/2015 0.73 0.00 6/7/2014 0.36 0.00 3/24/2015 0.74 0.00 6/8/2014 0.28 0.00 3/25/2015 0.76 0.00 6/9/2014 0.21 0.00 3/26/2015 0.76 0.00 6/10/2014 0.27 0.00 3/27/2015 0.76 0.00 6/11/2014 0.38 0.00 3/28/2015 0.79 0.00 6/12/2014 0.35 0.00 3/29/2015 0.78 0.00 6/13/2014 0.28 0.00 3/30/2015 0.82 0.00 6/14/2014 0.28 0.00 3/31/2015 0.83 0.00 6/15/2014 0.36 0.00 4/1/2015 0.83 0.00 6/16/2014 0.40 0.00 4/2/2015 0.86 0.44 6/17/2014 0.42 0.00 4/3/2015 0.89 0.00 6/18/2014 0.47 0.00 4/4/2015 0.86 0.00 6/19/2014 0.48 0.00 4/5/2015 0.89 0.00 6/20/2014 0.52 0.00 4/6/2015 0.90 0.00 6/21/2014 0.40 0.00 4/7/2015 0.90 0.00 6/22/2014 0.37 0.00 4/8/2015 0.93 0.00 6/23/2014 0.37 0.00 4/9/2015 0.93 0.00 6/24/2014 0.34 0.00 4/10/2015 0.90 0.00 6/25/2014 0.32 0.00 4/11/2015 0.90 0.00 6/26/2014 0.53 1.32 4/12/2015 0.91 0.00 6/27/2014 0.64 0.00 4/13/2015 0.88 0.00 6/28/2014 0.65 0.00 4/14/2015 0.86 0.00 6/29/2014 0.67 0.00 4/15/2015 0.84 0.00 6/30/2014 0.63 0.00 4/16/2015 0.79 0.00 7/1/2014 0.68 0.44 4/17/2015 0.79 0.00 7/2/2014 0.70 0.44 4/18/2015 0.83 0.00 7/3/2014 0.72 0.00 4/19/2015 0.83 0.00 7/4/2014 0.78 0.44 4/20/2015 0.84 0.00 7/5/2014 0.81 0.44 4/21/2015 0.85 0.00 7/6/2014 0.74 0.00 4/22/2015 0.90 0.00 7/7/2014 0.66 0.00 4/23/2015 0.85 0.00 7/8/2014 0.62 0.00 4/24/2015 0.85 0.00 7/9/2014 0.61 0.00 4/25/2015 0.82 0.00 7/10/2014 0.55 0.00 4/26/2015 0.72 0.00 7/11/2014 0.60 0.00 4/27/2015 0.73 0.00 7/12/2014 0.67 0.00 4/28/2015 0.76 0.00 7/13/2014 0.60 0.00 4/29/2015 0.76 0.00

7/14/2014 0.63 0.00 4/30/2015 0.73 0.00 7/15/2014 0.58 0.00 5/1/2015 0.78 0.00 7/16/2014 0.59 0.00 5/2/2015 0.83 0.00 7/17/2014 0.51 0.00 5/3/2015 0.84 0.00 7/18/2014 0.53 0.00 5/4/2015 0.70 0.00 7/19/2014 0.51 0.00 5/5/2015 0.66 0.00 7/20/2014 0.51 0.00 5/6/2015 0.65 0.00 7/21/2014 0.53 0.00 5/7/2015 0.64 0.00 7/22/2014 0.57 0.00 5/8/2015 0.65 0.00 7/23/2014 0.57 0.00 5/9/2015 0.67 0.00 7/24/2014 0.55 0.00 5/10/2015 0.66 0.00 7/25/2014 0.53 0.00 5/11/2015 0.66 0.00 7/26/2014 0.54 0.00 5/12/2015 0.73 0.00 7/27/2014 0.51 0.00 5/13/2015 0.76 0.00 7/28/2014 0.49 0.00 5/14/2015 0.82 0.00 7/29/2014 0.43 0.00 5/15/2015 0.79 0.00 7/30/2014 0.40 0.00 5/16/2015 0.79 0.00 7/31/2014 0.41 0.00 5/17/2015 0.78 0.00 8/1/2014 0.44 0.00 5/18/2015 0.65 0.00 8/2/2014 0.49 0.00 5/19/2015 0.69 0.00 8/3/2014 0.53 0.00 5/20/2015 0.70 0.00 8/4/2014 0.51 0.00 5/21/2015 0.74 0.00 8/5/2014 0.45 0.00 5/22/2015 0.78 0.00 8/6/2014 0.43 0.44 5/23/2015 0.73 0.00 8/7/2014 0.50 0.00 5/24/2015 0.80 0.00 8/8/2014 0.51 0.00 5/25/2015 0.80 0.00 8/9/2014 0.47 0.00 5/26/2015 0.81 0.00 8/10/2014 0.50 0.00 5/27/2015 0.81 0.00 8/11/2014 0.47 0.00 5/28/2015 0.80 0.00 8/12/2014 0.45 0.00 5/29/2015 0.80 0.00 8/13/2014 0.45 0.00 5/30/2015 0.74 0.00 8/14/2014 0.46 0.00 5/31/2015 0.73 0.00 8/15/2014 0.50 0.00 6/1/2015 0.74 0.00 8/16/2014 0.51 0.00 6/2/2015 0.70 0.00 8/17/2014 0.52 0.00 6/3/2015 0.69 0.00 8/18/2014 0.57 0.00 6/4/2015 0.76 0.00 8/19/2014 0.63 0.00 6/5/2015 0.75 0.00 8/20/2014 0.59 0.00 6/6/2015 0.71 0.00 8/21/2014 0.54 0.00 6/7/2015 0.65 0.00 8/22/2014 0.49 0.00 6/8/2015 0.70 0.00

8/23/2014 0.53 0.00 6/9/2015 0.75 0.00 8/24/2014 0.53 0.00 6/10/2015 0.72 0.00 8/25/2014 0.53 0.44 6/11/2015 0.72 0.00 8/26/2014 0.55 0.00 6/12/2015 0.70 0.88 8/27/2014 0.55 0.00 6/13/2015 0.73 0.00 8/28/2014 0.52 0.00 6/14/2015 0.71 0.00 8/29/2014 0.49 0.00 6/15/2015 0.59 0.00 8/30/2014 0.46 0.00 6/16/2015 0.56 0.00 8/31/2014 0.36 0.00 6/17/2015 0.63 0.00 9/1/2014 0.43 0.00 6/18/2015 0.64 0.00 9/2/2014 0.47 0.00 6/19/2015 0.64 0.00 9/3/2014 0.45 0.00 6/20/2015 0.67 0.00 9/4/2014 0.50 0.00 6/21/2015 0.63 0.00 9/5/2014 0.51 0.00 6/22/2015 0.58 0.00 9/6/2014 0.48 0.00 6/23/2015 0.56 0.00 9/7/2014 0.44 0.00 6/24/2015 0.67 0.00 9/8/2014 0.36 0.00 6/25/2015 0.62 0.00 9/9/2014 0.38 0.00 6/26/2015 0.65 0.00 9/10/2014 0.45 0.00 6/27/2015 0.64 0.00 9/11/2014 0.49 0.00 6/28/2015 0.62 0.00 9/12/2014 0.48 0.00 6/29/2015 0.57 0.00 9/13/2014 0.50 0.00 6/30/2015 0.57 0.00 9/14/2014 0.44 0.00 7/1/2015 0.69 0.00 9/15/2014 0.48 0.00 7/2/2015 0.73 0.00 9/16/2014 0.50 0.00 7/3/2015 0.76 0.00 9/17/2014 0.50 0.00 7/4/2015 0.77 0.00 9/18/2014 0.54 0.00 7/5/2015 0.69 0.00 9/19/2014 0.49 0.00 7/6/2015 0.70 0.00 9/20/2014 0.49 0.00 7/7/2015 0.72 0.00 9/21/2014 0.45 0.00 7/8/2015 0.68 0.00 9/22/2014 0.50 0.00 7/9/2015 0.63 0.00 9/23/2014 0.50 0.00 7/10/2015 0.60 0.00 9/24/2014 0.52 0.00 7/11/2015 0.55 0.00 9/25/2014 0.47 0.00 7/12/2015 0.64 0.00 9/26/2014 0.50 0.00 7/13/2015 0.65 0.00 9/27/2014 0.48 0.00 7/14/2015 0.68 0.00 9/28/2014 0.43 0.00 7/15/2015 0.64 0.00 9/29/2014 0.43 0.00 7/16/2015 0.63 0.00 9/30/2014 0.40 0.00 7/17/2015 0.65 0.00 10/1/2014 0.40 0.00 7/18/2015 0.63 0.00

10/2/2014 0.43 0.00 7/19/2015 0.67 0.00 10/3/2014 0.43 0.00 7/20/2015 0.75 0.00 10/4/2014 0.44 0.00 7/21/2015 0.77 0.00 10/5/2014 0.44 0.00 7/22/2015 0.76 0.00 10/6/2014 0.45 0.00 7/23/2015 0.80 0.00 10/7/2014 0.45 0.88 7/24/2015 0.79 0.00 10/8/2014 0.43 0.00 7/25/2015 0.72 0.00 10/9/2014 0.42 0.00 7/26/2015 0.70 0.00 10/10/2014 0.37 0.00 7/27/2015 0.70 0.00 10/11/2014 0.42 0.00 7/28/2015 0.69 0.44 10/12/2014 0.38 0.00 7/29/2015 0.79 0.00 10/13/2014 0.42 0.00 7/30/2015 0.74 0.00 10/14/2014 0.42 0.00 7/31/2015 0.71 0.00 10/15/2014 0.43 0.00 8/1/2015 0.72 0.00 10/16/2014 0.43 0.00 8/2/2015 0.62 0.00 10/17/2014 0.42 0.00 8/3/2015 0.60 0.00 10/18/2014 0.45 0.00

Appendix C Soil moisture dynamics and fog patterns at Gobabeb and Kleinberg.

Gobabeb Kleinberg Gravel plain Sand dune Gravel plain Time bare soil bare soil vegetation Fog bare soil Fog Depth 4 cm 4 cm 4 cm 5 cm Unit m^3/m^3 m^3/m^3 m^3/m^3 mm m^3/m^3 mm 8/19/2015 1.77 0.36 0.44 0.00 0.00 0.00 8/20/2015 1.68 0.34 0.37 0.00 0.00 6.00 8/21/2015 1.69 0.35 0.33 0.30 0.00 0.00 8/22/2015 1.83 0.33 0.33 0.60 0.00 2.20 8/23/2015 1.82 0.36 0.35 0.00 0.00 2.60 8/24/2015 1.65 0.36 0.35 0.00 0.00 0.00 8/25/2015 1.67 0.37 0.36 0.00 0.00 0.60 8/26/2015 2.00 0.40 0.43 0.40 0.00 0.50 8/27/2015 1.95 0.40 0.40 0.60 0.00 2.00 8/28/2015 1.80 0.40 0.40 0.60 0.00 3.90 8/29/2015 1.70 0.39 0.41 0.00 0.00 3.70 8/30/2015 2.01 0.40 0.42 3.10 0.00 1.10 8/31/2015 2.08 0.41 0.43 0.00 0.00 0.00 9/1/2015 1.78 0.43 0.43 0.00 0.00 11.40 9/2/2015 1.51 0.44 0.46 0.00 0.00 5.50 9/3/2015 1.31 0.45 0.50 0.00 0.00 2.20 9/4/2015 1.18 0.44 0.48 0.00 0.00 5.10 9/5/2015 1.25 0.42 0.44 0.00 0.00 1.60 9/6/2015 1.62 0.40 0.45 2.50 0.00 0.00 9/7/2015 1.60 0.40 0.44 0.00 0.00 0.00 9/8/2015 1.49 0.44 0.46 0.00 0.00 0.10 9/9/2015 1.48 0.43 0.46 2.50 0.00 0.00 9/10/2015 1.52 0.41 0.45 0.00 0.00 0.00 9/11/2015 1.43 0.50 0.50 0.00 0.00 0.00 9/12/2015 1.39 0.46 0.49 0.00 0.00 0.00 9/13/2015 1.49 0.48 0.49 0.00 0.00 4.60 9/14/2015 1.51 0.48 0.49 0.00 0.00 3.10 9/15/2015 1.54 0.47 0.50 0.10 0.00 2.70 9/16/2015 1.58 0.48 0.50 1.60 0.00 5.30 9/17/2015 1.88 0.47 0.49 3.40 0.00 0.50 9/18/2015 1.78 0.49 0.49 0.00 0.00 0.00 9/19/2015 1.62 0.45 0.48 1.00 0.00 0.00

9/20/2015 1.63 0.46 0.49 0.90 0.00 0.00 9/21/2015 1.58 0.45 0.48 0.00 0.00 1.20 9/22/2015 1.52 0.48 0.50 0.00 0.00 0.60 9/23/2015 1.48 0.48 0.49 0.00 0.00 0.60 9/24/2015 1.37 0.51 0.51 0.00 0.00 2.90 9/25/2015 1.30 0.53 0.54 0.00 0.00 0.90 9/26/2015 1.33 0.54 0.53 0.00 0.00 0.50 9/27/2015 1.50 0.54 0.53 0.00 0.00 0.50 9/28/2015 1.47 0.54 0.54 0.30 0.00 1.20 9/29/2015 1.49 0.50 0.53 0.00 0.00 0.00 9/30/2015 1.48 0.50 0.49 0.00 0.00 0.00 10/1/2015 1.47 0.53 0.52 0.00 0.00 0.40 10/2/2015 1.55 0.55 0.53 0.50 0.00 0.10 10/3/2015 1.54 0.55 0.54 2.10 0.00 0.60 10/4/2015 1.61 0.58 0.58 4.40 0.00 0.00 10/5/2015 1.57 0.63 0.62 0.00 0.00 0.20 10/6/2015 1.37 0.65 0.63 0.00 0.00 0.00 10/7/2015 1.40 0.65 0.61 0.00 0.00 0.00 10/8/2015 1.44 0.61 0.60 0.00 0.00 1.30 10/9/2015 1.53 0.61 0.59 2.30 0.00 0.40 10/10/2015 1.60 0.60 0.60 0.20 0.00 0.20 10/11/2015 1.45 0.58 0.60 0.10 0.00 0.00 10/12/2015 1.52 0.54 0.58 0.00 0.00 0.00 10/13/2015 1.57 0.55 0.58 0.00 0.00 0.00 10/14/2015 1.44 0.55 0.57 0.00 0.00 0.00 10/15/2015 1.38 0.57 0.58 0.00 0.00 0.00 10/16/2015 1.32 0.58 0.60 1.00 0.00 1.10 10/17/2015 1.47 0.58 0.60 4.40 0.00 0.00 10/18/2015 1.48 0.58 0.60 0.00 0.00 0.40 10/19/2015 1.43 0.62 0.64 0.00 0.00 4.30 10/20/2015 1.43 0.64 0.66 0.00 0.00 3.60 10/21/2015 1.44 0.61 0.64 0.00 0.00 0.00 10/22/2015 1.45 0.62 0.66 0.00 0.00 0.20 10/23/2015 1.48 0.64 0.66 0.00 0.00 0.60 10/24/2015 1.67 0.62 0.64 0.00 0.00 0.60 10/25/2015 1.69 0.63 0.65 0.00 0.00 0.00 10/26/2015 1.48 0.65 0.67 0.00 0.00 0.00 10/27/2015 1.45 0.65 0.68 0.00 0.00 0.00 10/28/2015 1.54 0.68 0.68 0.00 0.00 0.30 10/29/2015 1.53 0.65 0.66 0.00 0.00 2.40

10/30/2015 1.58 0.61 0.63 0.00 0.00 0.00 10/31/2015 1.69 0.61 0.64 2.80 0.00 0.00 11/1/2015 1.63 0.59 0.63 0.50 0.00 0.00 11/2/2015 1.44 0.61 0.62 0.00 0.00 0.00 11/3/2015 1.31 0.50 0.68 0.00 0.00 0.00 11/4/2015 1.38 0.30 0.70 0.00 0.00 0.00 11/5/2015 1.59 0.43 0.71 0.00 0.00 0.00 11/6/2015 1.78 0.57 0.75 0.00 0.00 0.00

Curriculum Vitae

Bonan Li

Education

Master of Science

Indiana University – Purdue University Indianapolis (IUPUI),

Indiana, USA, August 2017

Focusing on exploring the impact of rainfall and non – rainfall waters on soil moisture dynamics in the Namib Desert.

Bachelor of Science

Chengdu University of Technology (CDUT),

Chengdu, China, June 2014

Major in Applied physics with focus on using Landsat imagery to detect algal blooms in the Reservoirs/Lakes of Songnen Plain.

Experiences

Teaching Assistant, Environmental Geology Lab, IUPUI, Department of Earth Sciences,

Indianapolis, Indiana (August 2016 – May 2017).

Research Assistant, IUPUI, Department of Earth Sciences, Indianapolis, Indiana (August

2015 – June 2016).

Research associate, Northeast Institute of Geography and Agroecology (IGA), Chinese

Academy of Sciences (CAS), Changchun, China (June 2014 – June 2015).

Presentations

Li B., L. Wang, K. Kaseke, L. Li, M. Seely. (2016). The impact of rainfall and its control in a foggy desert. Poster. The American Geophysical Union (AGU), Fall Meeting, San

Francisco.

Manuscripts

Remote extraction of algal bloom in the reservoirs and lakes of Songnen Plain usng Landsat imagery (Completed).

Challenge for irrigation based agriculture in China: Water Scarcity. (A report for The U. S.

State Department’s Office of Conservation and Water, Bureau of Oceans Environment and

Sciences (OES/ECW))

Publications

Li, B., L. Wang, K. Kaseke, L. Li. The impact of fog on soil moisture dynamics in the

Namib Desert. Advanced in Water Resources, submitted.

Li B., L. Wang, K. Kaseke, L. Li, M. Seely. 2016. The impact of rainfall on soil moisture dynamics in a foggy desert. PLoS ONE 11(10).

Lu, X., Wang, L., Ming, P., Kaseke, K., and Li, B. 2016. A multi-scale analysis of Namibian rainfall over the recent decade - comparing TMPA satellite estimates and ground observations. Journal of Hydrology Regional Studies 8: 59-68.