Environmental differences in tropical soil temperatures in Kenya

Environmental temperature differences within tropical soils are a function of the total solar radiation received at the surface, and also depend on woody vegetation cover. Environmental soil temperature is recorded by carbon isotope substitutions during the formation of calcite, which is preserved in paleosols. Therefore, analysis of preserved carbonates can be used as a proxy indicator of paleotemperatures. Preliminary data from hominid sites in the Turkana Basin show that soil temperatures have been in excess of 30°C for much of the past 4-6 million years in that region. In this study two years of continual subsurface soil monitoring were conducted at 28 sites within and around Kenyan National Parks and we present annual and seasonal averages of soil temperatures at a depth of 25 cm within different microclimates, which should approximate the absolute formation temperature of soil carbonates in present day tropical soils. In addition, we use an iterative method to solve the heat diffusion equation to estimate the soil surface temperature. In the tropics, where the solar angle is high throughout the year, observed environmental temperature differences over small spatial distances are as high as 30°C in the most extreme contrasts between grassland and forested microclimates. Average soil temperatures at 25 cm depth are highest in the Turkana basin where annual and seasonal averages are in excess of 30°C. These results are consistent with the paleotemperature measurements, indicating that temperatures are as hot today as they have been over the past several million years.

ENVIRONMENTAL DIFFERENCES IN TROPICAL SOIL TEMPERATURES IN KENYA by William Davis Mace A thesis submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Master of Science in Geology Department of Geology and Geophysics The University of Utah December 2012 Copyright © William Davis Mace 2012 All Rights Reserved The Uni v e r s i t y of Utah Graduat e School STATEMENT OF THESIS APPROVAL The thesis of William Davis Mace has been approved by the following supervisory committee members: Thure E. Cerling______________ , Chair 6 /1 /1 2 Date Approved Francis H. Brown______________ , Member 6 /1 /1 2 Date Approved David S. Chapman_____________ , Member 6 /1 /1 2 Date Approved and by __________________ Douglas K. Solomon__________________ , Chair of the Department of __________________Geology and Geophysics_______________ and by Charles A. Wight, Dean of The Graduate School. ABSTRACT Environmental temperature differences within tropical soils are a function of the total solar radiation received at the surface, and also depend on woody vegetation cover. Environmental soil temperature is recorded by carbon isotope substitutions during the formation of calcite, which is preserved in paleosols. Therefore, analysis of preserved carbonates can be used as a proxy indicator of paleotemperatures. Preliminary data from hominid sites in the Turkana Basin show that soil temperatures have been in excess of 30°C for much of the past 4-6 million years in that region. In this study two years of continual subsurface soil monitoring were conducted at 28 sites within and around Kenyan National Parks and we present annual and seasonal averages of soil temperatures at a depth of 25 cm within different microclimates, which should approximate the absolute formation temperature of soil carbonates in present day tropical soils. In addition, we use an iterative method to solve the heat diffusion equation to estimate the soil surface temperature. In the tropics, where the solar angle is high throughout the year, observed environmental temperature differences over small spatial distances are as high as 30°C in the most extreme contrasts between grassland and forested microclimates. Average soil temperatures at 25 cm depth are highest in the Turkana basin where annual and seasonal averages are in excess of 30°C. These results are consistent with the paleotemperature measurements, indicating that temperatures are as hot today as they have been over the past several million years. TABLE OF CONTENTS ABSTRACT.......................................................................................................................iii LIST OF TABLES............................................................................................................. vi LIST OF FIGURES.......................................................................................................... vii Chapters 1. INTRODUCTION........................................................................................................ 1 Kenya Meteorology................................................................................................ 5 References............................................................................................................... 8 2. SPATIAL AND TEMPORAL TROPICAL SOIL TEMPERATURES.................10 Abstract..................................................................................................................10 Introduction........................................................................................................... 11 Methods and Materials........................................................................................ 13 Site selection and temperature measurement............................................. 13 Soil temperature data collection.................................................................. 15 Quantification of site specific canopy coverage........................................16 Soil data analysis.......................................................................................... 16 Results....................................................................................................................17 Discussion ............................................................................................................. 28 Soil temperature relationship to local ecology and climate......................28 Soil temperature relationship to carbonate formation............................... 35 Land use effects on soil temperature...........................................................38 Conclusions........................................................................................................... 39 References............................................................................................................. 44 3. TROPICAL ENVIRONMENTAL SURFACE TEMPERATURES PRESENTED AS A COMPOSITE DAY......................................................................................... 46 Abstract .................................................................................................................. 46 Introduction........................................................................................................... 47 Methods and Materials ........................................................................................ 49 Soil temperature data collection.................................................................. 49 Site selection and description...................................................................... 49 Description of the seasons............................................................................53 Data Analysis Procedure...................................................................................... 53 Composite Day.....................................................................................................56 Results....................................................................................................................60 Discussion............................................................................................................. 66 Validation of high soil surface temperatures............................................. 66 Air temperature and soil temperature comparison.....................................67 Relationship to Human Heat Tolerance..............................................................73 Conclusions........................................................................................................... 75 References............................................................................................................. 77 4. SUMMARY AND CONCLUSIONS....................................................................... 79 Appendices A. SITE LOCATION AND DESCRIPTION................................................................81 B. ILERET AND TURKWEL WEATHER STATIONS: THE FIST YEAR OF DATA....................................................................................87 C. ADDITIONAL TWO YEAR DATASETS............................................................104 D. ADDITIONAL COMPOSITE DAY FIGURES................................................... 110 v LIST OF TABLES 2.1 Daily average computed using two different methods............................................ 18 2.2 Canopy coverage, mean annual, seasonal average, and soil temperature anomaly results for soil temperature at 25 cm depth for June 2009 to June 2011................19 2.3 Elevation corrected mean air temperature and precipitation measurements from nearby climate monitoring sites.................................................................................22 3.1 Latitude, longitude, and altitude of surface temperature sites.................................51 3.2 Values for diffusivity (a#) for selected soil temperature sites............................... 58 3.3 Annual and seasonal mean daily maximum, minimum, and ranges in both soil surface and air temperatures for Nairobi National Park, Tana River Primate Reserve, and Meru National Park..............................................................................................71 A.1 Latitude, longitude, and elevation of soil temperature sites....................................83 B.1 Parts list for the weather stations at Ileret and Turkwel field stations...................90 B.2 Latitude, Longitude, and elevation of the weather stations.....................................91 LIST OF FIGURES 2.1 Digital Elevation Model (DEM) of Kenya with soil temperature sites (green stars) and weather station sites (red stars)............................................14 2.2 One year of data from June 1, 2009 until May 20, 2010 at the Ileret Field station of the Turkana Basin Institute measured at 0.25 m depth........................................................................................................................25 2.3 Plotted are the mean annual soil temperatures at 0.25 m depth with standard deviation plotted on the top of each bar .................................26 2.4 Seasonal average soil temperatures at 0.25 m .........................................................27 2.5 Seasonal averaged 2 m air temperatures from nearby weather stations.................29 2.6 Seasonal precipitation amounts at the weather stations near the measurement sites.......................................................................................................30 2.7 Seasonal temperature anomalies (mean annual soil temperature difference from mean annual air temperature)..........................................................31 2.8 Mean annual soil temperature and elevation ...........................................................33 2.9 The relationship of canopy cover to the annual mean soil temperature at 25 cm normalized to zero elevation................................................. 34 2.10 Daily averaged temperature at a depth of 0.25 m from the different microclimatological settings from four different national parks and reserves.....37 2.11 Total land area in Kenya that has been designated for agriculture and population during the past 60 years.................................................................. 41 2.12 Seasonal averaged soil temperatures at all four Meru vegetation settings ......... 42 2.13 Mean annual temperatures at Arabuko Sokoke National Forest.......................... 43 3.1 Digital elevation model showing the locations of the study sites........................... 52 3.2 RMS Error as a function of iteration number of the diffused soil temperature signal compared with the observed soil temperature signal at 5 cm depth...............................................................................57 3.3 One month of modeled data from the Tana River Primate Reserve open s ite ...... 59 3.4 One year of modeled surface temperature for the different environmental types at Meru National Park....................................................................................... 63 3.5 One year of modeled surface temperature for the different environmental types at Tana River Primate Reserve (TRPR).................................64 3.6 One year of modeled surface temperature for the different environmental types at Nairobi National Park..........................................................65 3.7 One day of data comparing the accuracy of a soil temperature sensor that is exposed directly to the sun versus a sensor that is covered by 1-2 mm of soil.......................................................................................... 68 3.8 Measured daily maximum surface temperatures at TBI Turkwel station from October 2011 to November 2011.........................................................69 3.9 Comparison of composite days composed from the test period October 2011 to November 2011at TBI Turkwel Station................................................................... 74 B.1 Annotated image of the weather station that is located at Turkwel Field Station.................................................................................................................92 B.2 Daily air temperature high, low, range, and average at Ileret on the east side of Lake Turkana from June 2010 through May 2011.............................................. 93 B.3 Daily air temperature high, low, range, and average at the Turkwel weather station on the west side of Lake Turkana from June 2010 through May 2011....94 B.4 Soil moisture and precipitation measured at the Ileret weather station from June 2010 through May 2011............................................................................97 B.5 Soil moisture and precipitation measured at the Turkwel weather station during June 2010 through May 2011.........................................................................98 B.6 Soil temperature and precipitation at the Ileret weather station measured June 2010 through May 2011................................................................... 99 B.7 Soil temperature and precipitation at the Turkwel weather station measured June 2010 through May 2011................................................................. 100 B.8 Daily averaged solar insolation measured at the Ileret weather station viii during June 2010 through May 2011...................................................................... 101 B.9 Daily averaged solar insolation measured at the Turkwel weather station during June 2010 through May 2011...................................................................... 102 C.1 Daily average temperature for Nakuru National Park for two years, from June 2009 until May 2011........................................................................................ 105 C.2 Daily average temperature for Mt. Kenya National Park for two years, from June 2009 until May 2011........................................................................................ 106 C.3 Daily average temperature for Kakamega Forest National Park for two years, from June 2009 until May 2011...............................................................................107 C.4 Daily average temperature for Shimba Hills National Park for two years, from June 2009 until May 2011........................................................................................ 108 C.5 Daily average temperature for Tsavo West National Park for one year, from June 2009 until May 2010........................................................................................ 109 D.1 Composite day for Arobuko Sokoke National Park from June 2009 to May 2010.......................................................................................... 111 D.2 Composite day for Arobuko Sokoke National Park from June 2010 to May 2011.......................................................................................... 112 D.3 Composite day for Mt. Kenya National Park from June 2009 to May 2010.......................................................................................... 113 D.4 Composite day for Nakuru National Park from June 2009 to May 2010.......................................................................................... 114 D.5 Composite day for Nakuru National Park from June 2010 to May 2011.......................................................................................... 115 D.6 Composite day for Shimba Hills National Park from June 2009 to May 2010.......................................................................................... 116 ix CHAPTER 1 INTRODUCTION Earth's dynamic climate system naturally experiences gradual changes over time. However, during the last century, since the dawn of the modern industrial age, Earth's climate has been changing at a much more accelerated rate. It is widely accepted that human activities have been amplifying changes in global temperature through burning large amounts of fossil fuels. The product of the combustion of fossil fuels, carbon dioxide, is a greenhouse gas that alters atmospheric transparency making it more opaque to long wave radiation emitted from the Earth's surface into space. The result is the inability of the surface to radiate the heat necessary to maintain a balanced surface energy budget which leads to an increase in the globally averaged surface air temperature and therefore, continued climate change [IPCC, 2007]. In addition to atmospheric emissions of green house gases, humans have also been altering the way that the Earth's surface absorbs energy from the sun through changes in land use practices [Chapin, 2002]. Altering the landscape through cities, deforestation, and livestock grazing, change how the Earth absorbs and reflects solar radiation. The amount of radiation reflected back into space by the surface is known as the albedo. Naturally the Earth has an average albedo of about 0.3 [Wallace and Hobbs, 2006]. The difference in albedo between a forest canopy and grassland is significant; 0.08 and 0.30, respectively, which has a major influence on the amount of energy available within an ecosystem [Chapin, 2002; Robert et al, 2008; Vitousek, 1994]. Together these changes, the greenhouse effect and land use practices, translate to an increase in the global temperature and ultimately to global climate change. During the last century, globally averaged surface temperature has risen by approximately 0.75 °C. Computer model simulations predict that in the coming decades, globally averaged surface temperatures will increase between 1-6°C [IPCC, 2007]. The magnitude of observable change will vary greatly across different regions of the planet and furthermore, the magnitude of changes will be different for each of the many ecosystems within a given landscape. Ecosystems are governed by the local climate, soil parent material, topography, potential biota, and time. These five independent control variables are called state factors and were explained in depth by Hans Jenny [Jenny, 1980]. Together, these five state factors set the bounds for the characteristics of a particular ecosystem. Ecosystem characteristics such as air temperature, canopy cover, and soil temperature are important controls in a variety of ecological processes and are also good indicators of climate change and the magnitude of changes within an ecosystem as they are a direct measure of the overall climate encompassing a particular ecosystem [Chapin, 2002; Jenny, 1980]. Processes such as plant growth and nutrient cycling are crucial components to the overall health of an ecosystem and are highly dependent on environmental controls such as the ones mentioned above [Vitousek, 1994]. Therefore, quantifying and continually monitoring these parameters is increasingly important in light of global climate and land 2 use changes. Monitoring these parameters will enable us to identify and further quantify changes within natural environments. In addition to present day applications these environmental characterizations, when combined with paleothermometry and carbonate isotopic analysis of paleosols to determine woody vegetation cover, can provide significant insight into the environmental conditions existing in specific areas over the past several million years. Furthermore, this information should be applicable to questions pertaining to the environmental context of human evolution. There has been a long-standing debate about environmental conditions that existed during critical periods of human evolution. It is widely accepted, however, that the last common ancestor (LCA) between humans and chimpanzees lived in a wooded forest environment. Since the LCA 5 to 8 million years ago (mya), as the two species diverged, it is thought that habitats became less wooded and more seasonally arid [Cerling et al., 2010; Potts, 1998; White et al., 2009; Wood and Harrison, 2011]. Increased aridity and grassland expansion is thought to have driven the further development of bipedalism and the reduction of body hair in hominins [Wheeler, 1984, 1991, 1992, 1993]. The environment(s) in which the genus Homo evolved undoubtedly provided shade, shelter, and food. Just as importantly, this environment, or the environmental mosaic of the landscape, would have had a major influence on the evolutionary development and adaptations of the genus through time. The following thesis is divided into two principal chapters. In Chapter 2, I discuss the delta 47 (A47) paleothermometer, and present deep (0.25 m) soil temperature data from modern grassland, bushland and forest environments to show that the 3 seemingly high paleotemperatures calculated using this method are still observed within today's soil environments. Chapter 3 presents a composite day of soil temperatures from grassland, bushland, and forested environments throughout Kenya. The aim of this chapter is to quantify the average daily soil temperature cycles observed throughout Kenya's environmental mosaic. Results from this chapter can be applied to the study of hominid energy balance and provide insight on environmental factors driving hominid evolution. Data presented in both chapters were collected in present day environments within the Kenyan Wildlife Management System. The study locations were chosen within and near the Kenyan National Wildlife Reserves because of the relatively undisturbed nature and limited anthropogenic influences. Areas on the peripheries of the reserves have been influenced heavily by people and their animals and provide valuable insight into the effects of land use change. Because of the loose usage of the terms grassland, woodland, bushland, forest, etc. there is a need for these terms to be strictly defined early in this paper. For these definitions I adopt the definitions of such landscape types defined by the United Nations Educational, Science, and Cultural Organization [White, 1983] as follows. • "Forest. A continuous stand of trees at least 10-m tall, their crowns interlocking." • "Woodland / Bushland / Shrubland. Woodland: An open-stand of trees at least 8­m tall with a canopy cover of 40 % or more, and a field layer dominated by grasses." • "Bushland: An open-stand of bushes usually between 3- and 8-m tall with a canopy cover of 40 % or more." 4 • "Grassland. Land covered with grasses and other herbs, either without woody plants or the latter not covering more than 10 % of the ground." • "Desert. Arid landscapes with a sparse plant cover. The sandy, stony or rocky substrate contributes more to the appearance of the landscape than does the vegetation." Kenya Meteorology Kenya is located in equatorial Africa. This area spans more than 1.5 million km and is considered to be one of the more meteorologically complex parts of Africa [Griffiths, 1972; Nicholson, 1996]. Climates within this zone are extremely variable from hot humid coasts with over 2000 mm precipitation per year to hot dry scrubland with as little as 200 mm of precipitation per year to dense tropical forests in cool, wet highlands. Topography in Kenya is just about as diverse as climate. Kenya lies atop a geologically active zone (the Kenya Dome) that is bisected by a central rift valley. On the eastern and western margins of this rift lie the Kenyan highlands with elevations between 1700 and 2100 m on broad flat plains interrupted by mountains of still higher elevation. Southeastern Kenya is mainly flat and is < 300 m above sea level. Most of northern Kenya lies below 1000 m elevation, but small highland area also exist. Elevations range from sea level to glaciated mountain peaks above 5100 m. There are six main meteorological features that have a large influence on Kenyan meteorology throughout the year. These include three air streams and three convergence zones. The northwest and southeast monsoon are the result of the migration of the Inter- Tropical Convergence Zone (ITCZ). Both the northeast and southeast monsoons are dry 5 and affect the Kenyan region during the Northern Hemisphere winter and summer, respectively. The Kenyan "long rains" and the "short rains" occur during the migration periods of the ITCZ, which occur during the Northern Hemisphere spring and fall, respectively. The third air stream that influences the Kenyan region is the Congo airstream with its westerlie and southwesterlie components. The boundary between the westerly and southwesterly components of the Congo air mass, the Congo Air Boundary, is situated perpendicular to and south of the ITCZ. The monsoons, unlike the Asian monsoon, are convectively stable and therefore, are associated with dry conditions. Humid air flowing out of the Congo Basin is convectively unstable and therefore, associated with rain during the transition seasons [Nicholson, 1996]. These atmospheric features along with widely variable topography lead to an extremely complex distribution in precipitation and temperature both temporally and geographically [Griffiths, 1972; Levin, 2008 Nicholson, 1996]. Climatological season is defined by the Kenya Meteorology Department: December, January, and February (DJF) are the "warm and dry season," March, April, and May (MAM) are the "long rains," June, July, and August (JJA) are the "cool and dry season," and September, October, and November (SON) are the "short rains." Geographically, the distribution of rainfall is more straightforward. Rainfall over Kenya is driest in the north and northeastern lowlands and wettest in the southwest and costal southeast. In the driest parts of Kenya, the Turkana basin, mean annual precipitation (MAP) values are less than 200 mm per year and in the wettest parts of Kenya, in the Kenyan Highlands and around Mt. Kenya, MAP values are greater than 2500 mm. 6 The seasonal distribution of temperature over Kenya is such that, typically, the coldest month recorded is in July or August and the warmest month is November or December [Griffiths, 1972]. Seasonal temperature fluctuations through the region are small if existent at all because of the close proximity to the equator, however, there are observable differences in temperature due to the rainy seasons, but these have very little to do with the small variability in solar insolation. Temperature in Kenya is not only dependent on solar insolation, but also on the elevation of a given location. The dependence of temperature with elevation is known as the environmental lapse rate and is typically measured in units of C° of temperature drop per km of elevation gain. Therefore, it can be expected that the hottest regions of Kenya will be those that are low in elevation and the coolest will be those that are highest in elevation. 7 8 References Cerling, T. E., N. E. Levin, J. Quade, J. G. Wynn, D. L. Fox, J. D. Kingston, R. G. Klein, and F. H. Brown (2010), Comment on the paleoenvironment of Ardipithecus ramidus, Science 328(5982): 1105. Chapin, F. S., P. A. Matson, and H. A. Mooney (2002), Principles o f Terrestrial Ecosystem Ecology, Springer, New York. Griffiths, J. F. (1972), Climates o f Africa, Elsevier Publishing Co., Amsterdam, New York. IPCC, 2007: Climate Change 2007: The Physical Science Basis. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)], Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jenny, H. (1980), The Soil Resource, Springer, Heidelberg, Berlin. Levin, N. E. (2008), Isotopic Records of the Plio-Pleistocene Climate and Environments in Eastern Africa. PhD dissertation, Dept. of Geol. and Geophys., Univ. of Utah Salt Lake City, Utah. Nicholson, S. E. (1996), A review o f Climate Dynamics and Climate Variability in Eastern Africa, The Limnology, Climatology and Paleoclimatology o f the East African Lakes, pp. 25-56, edited by T. Johnson, E.B Odada, K. T. Whittaker, Gordon and Breach Science Publishers, Amsterdam, The Netherlands. Potts, R. (1998), Environmental hypotheses of hominin evolution, Yearbook o f Physical Anthropology 41: 93-136. Robert, B. J., and T. R. James, et al. (2008). Protecting climate with forests, Environmental Research Letters 3(4): 044006. Vitousek, P. M. (1994), Beyond global warming: Ecology and global change, Ecology 75(7): 1861-1876. Wallace, J. M. and P. V. Hobbs (2006), Atmospheric Science: An Introductory Survey, Elsevier Academic Press, New York. Wheeler, P. E. (1984), The evolution of bipedality and loss of functional body hair in hominids, Journal o f Human Evolution, 13: 91-98. 9 Wheeler, P. E. (1991a), The thermoregulatory advantages of hominid bipedalism in open equatorial environments: The contribution of increased convective heat loss and cutaneous evaporative cooling, Journal o f Human Evolution, 21: 107-115. Wheeler, P. E. (1991b), The influence of bipedalism on the energy and water budgets of early hominids, Journal o f Human Evolution, 21: 117-136. Wheeler, P. E. (1992a), The thermoregulatory advantages of large body size for hominids foraging in savannah environments, Journal o f Human Evolution, 23: 351-362. Wheeler, P. E. (1992b), The influence of the loss of functional body hair on the water budgets of early hominids, Journal o f Human Evolution, 23: 379-388. Wheeler, P. E. (1993), The influence of stature and body form on hominid energy and water budgets; a comparison of Australopithecus and early Homo physiques, Journal o f Human Evolution, 24: 13-28. White, F. (1983), The vegetation of Africa, Natural Resources Research, 20, 1-356 White, T. D., B. Asfaw, et al. (2009), Ardipithecus ramidus and the paleobiology of early hominids, Science, 326(5949): 64, 75-86. Wood, B. and T. Harrison (2011), The evolutionary context of the first hominins, Nature 470(7334): 347-352. CHAPTER 2 SPATIAL AND TEMPORAL TROPICAL SOIL TEMPERATURES Abstract Temperatures within the Turkana Basin are hot. Preliminary soil carbonate formation temperatures, (A47) paleothermometry data, from hominid sites within the Turkana Basin suggest that the absolute temperature of the soil during carbonate formation has been in excess of 30°C during much of the past 4-6 million years. In the following chapter we present soil temperature data from modern soils at 28 different sites located in national parks and wildlife refuges throughout Kenya at depths consistent with soil carbonate formation. The data presented show that soil temperatures within the Turkana Basin and at the Tana River Primate Reserve are hotter than any other area within Kenya on both seasonal and annual timescales. Average temperatures on both time scales exceed 30°C. We also present climate data collected from nearby weather stations to associate the local climate and vegetation conditions and the effects these variables have on deep soil temperatures. Introduction Within the paleoclimate community and the archeological community there is a need to understand the evolution of the landscape over time, particularly temperatures that may have been experienced by the present and past residents of these landscapes such as hominids [Cerling, 2011; Cerling et al., 2010; White et a l, 2010]. Over the past few decades, different methods have been developed as proxy thermometers to calculate the absolute soil temperatures that were present during pedogenic carbonate formation using stable isotopes [Eiler, 2007]. One method, the A47 paleothermometer [Ghosh et al., 2006] uses the rare 13C-18O bond in carbonate minerals for such paleothermometry. Carbonate minerals are preserved within paleosols and, when extracted and isotopically analyzed, provide a measure of the temperature of the soil environment during the formation of the carbonate. When this method of paleothermometry is applied to the hominid sites in the Turkana Basin in northern Kenya, preliminary data reveal that paleosol temperatures have been in excess of 30°C during much of the past 4-6 mya [Passey et al., 2009]. Furthermore, it is commonly assumed that the mean annual soil temperature (MAST) will average 1-3° warmer than the mean annual air temperature (MAT) [Geiger, 1965]. This means that the air temperatures that are associated with the calculated soil temperatures in the sedimentary record are higher than mean annual air temperatures observed today and leaves open a question weather modern environments exist that produce soil temperatures, at the depths of carbonate precipitation, greater than or equal to 35°C. Because of this, it is useful to provide modern day analogs to the environments and microclimates that may have produced such high soil temperatures in the past. 11 Because pedogenic carbonates form at depths between .30 and .50 m in the soil column [Cerling, 1984; Passey et al., 2009] the temperature recorded during their formation represents a seasonal averaged representation of soil temperature, and most likely a warm season biased seasonal average [Breecker et al., 2009; Passey et al., 2009]. Processes that occur at the soil surface, such as solar heating and infrared cooling (net radiation), evaporative cooling, and sensible heat flux, which are components of the surface energy budget, that are coupled to local weather tend to control the soil temperature at depths within the soil column. Perturbations in the surface energy budget, which are not quantified here, are conducted through the soil column at a rate that is dependent on the soil type and water content. At any depth between the surface and a few meters, the daily average temperature is similar. However, the amplitude is increasingly attenuated and phase shifted with depth. At 0.25 m depth, the daily range is drastically attenuated leaving only seasonal and longer timescale transitions resolvable, thereby providing a more accurate measure of the long-term average soil temperature [Jury and Horton 2004; Lal and Shukla, 2004]. The average in soil temperature, which is representative of the carbonate formation temperature, can then be correlated with averages of the surface climate and surface characteristics, thus providing a connection between the surface characteristics and the carbonate formation temperature. Here we address the need for these environmental analogs to correlate with paleothermometry data by looking at modern day tropical environments located throughout the Kenyan National Park System. In doing so, the deep (0.25 m) soil temperatures will be presented on both annual and seasonal timescales using soil temperature records collected during two years at different sites throughout Kenyan 12 National Parks. To assess the effect that microclimates have on soil temperature, temperature sensors were placed in the soils of different ecological settings (e.g., forest, grasslands, and bushland) within each regional setting and left undisturbed for approximately two years. In addition to the deep soil temperature, we will also present some environmental characteristics of the area including regional air temperature, regional precipitation, type and density of vegetation cover. Methods and Materials Site selection and temperature measurement Twenty-eight sites in 11 regional localities in Kenya were studied to characterize soil temperatures relevant to soil carbonate formation (Figure 2.1). Soil temperature sites were chosen to represent a variety of climates and ecosystems in order to characterize the range of environments that may be relevant to the geological record of sedimentary environments. Most of the sites were within Kenyan National Parks or National Reserves and thus, have minimal human disturbance. In addition, a few sites on the peripheries of the parks were monitored to provide insight into the effects of land use changes by recent human activities. Within each regional locality several individual sites were chosen to assess the effect that woody cover and shade have on the soil temperature. Because of the loose usage of the term "savanna" [Ratnam et al., 2011] the definitions of the United Nations Educational, Science, and Cultural Organization [White, 1983] are used and are described in Appendix A. 13 14 Figure 2.1 - Digital Elevation Model (DEM) of Kenya with soil temperature sites (green stars) and weather station sites (red stars). Climate data were obtained from the Kenya Meteorological Service (KMS) using the operational weather station nearest each study site; in most cases this is within about 20 km of the field site. However, at some sites there were no nearby weather stations, so in these situations the station that was closest and judged to be most representative of the local climate was used. The elevation differences between the soil temperature sites and the climate monitoring sites were compensated for by adjusting the measured air temperatures of the climate stations based on an average environmental lapse rate (AELR) of -6.5°C per km of elevation change [Wallace, 2006]. No corrections were made for precipitation because variation in precipitation cannot be directly tied to elevation; the values reported at the KMS climate sites are used without modification. Soil temperature data collection Soil temperature measurements were collected using HOBO Pendent® data loggers. These have a stated measurement range of -20°C - +70°C, with an accuracy of ±0.5°C and a precision of 0.1°C. At each site the temperature loggers were buried in a 0.1 m wide trench at three depths: 0.05, 0.15, and 0.25 m. After the sensors were placed, the trench was back filled with the same soil material that was removed. Where possible, three distinct environmental types, forest, grassland, and bushland environments, as described by the UNESCO classification scheme, were targeted. In two cases, at Meru National Park and Arabuko Sokoke, disturbed areas related to land use change were also monitored. The loggers were programmed to record an instantaneous temperature measurement once every 20 minutes. The soils were monitored between May 2009 and May 2011. 15 Quantification of site specific canopy coverage For this study we define canopy coverage as the total area of tree crown visible directly overhead at each instrumented site. To quantify this variable we employed a 180° circular fisheye photograph. At each soil temperature site an image of the overlying tree cover was taken looking vertically upward from ~1 m above the ground using a 10.2 Mp Nikon D200 digital camera fitted with a Sigma f2.8 180° hemispherical fisheye lens. Image J software was used to crop the images from a full 180° so that only the vertical 60° (30° off the center point) of the image was considered. Adobe Photoshop was used to define the edge of the tree crowns using its enhanced edge finding capabilities. Image pixels that were classified as tree crown were colored black and those that did not contain tree crown, tree crown gap, were colored white. Software was written in the Interactive Data Language (IDL) programming language to import the classified images and count the number of pixels classified as tree crown. The percentage of canopy cover is defined as the percentage of pixels within the image that are classified as tree crowns. Soil data analysis The first step in the data analysis procedure was to trim the dataset so that only full days were considered. The first three days in the time series were removed to allow for the soil column to re-equilibrate after the installation of the temperature sensors, and the day that the instrument was retrieved was also removed from the time series. To remove outliers, the trimmed, raw data were then filtered by removing data points more 16 than three standard deviations from the mean. Since the loggers were undisturbed during the measurement period filtering yielded no change to the time series. The data sets were run through a series of codes written in the IDL to extract the daily high, daily low, and the daily average temperature at each depth. The daily average temperature was obtained using the average of the daily high and the daily low temperatures, which is the method that is consistent with the KMO. The other alternative is to calculate a running average of each measurement over the day; however, damping of the diurnal signal by the soil column makes the difference between the two methods negligible as shown in Table 2.1. Next, the annual and seasonal mean temperatures were calculated by computing the average of the data sets (annual average) and the three month sections (seasonal average) of the datasets. The code written for this task is presented in Appendix B. Results Table 2.2 shows: the mean annual soil temperature and standard deviation calculated from the daily mean at 0.25 m for June 2009 through June 2011, Tmast and aMAST, respectively; the seasonal 0.25 m soil temperature average and standard deviation calculated from the daily mean, T„Season„ and o „Season„, respectively; percentage canopy cover directly overhead at each site; and the annual difference between mean annual soil temperature and the mean annual air temperature, Tmast - Tair referred to as the anomaly. Table 2.3 shows the elevation-corrected seasonal and mean annual air temperature and precipitation averages. 17 Tabic 2.1 - Daily average computed using two different methods. The data in this table arc from the Tana River Primate Reserve Open site and were recorded on December 16, 2009. Average Method Daily Average (°C) Max+Min/2 34.55 Average of 20 Minute Samples 34.51 Tables 2.2 - Canopy coverage, mean annual, seasonal average, and soil temperature anomaly results for soil temperatures at 25 cm depth for June 2009 - June 2011. Site Canopy Cover (%) T/MAST CC) °MAS1' tjja C Q °JJA SON (*c) SON tdjf C Q °DJF tmam CC) aMAM T1 MAST Arabuko Brachystegia Forest 92 28.3 1.6 26.3 0.8 27.9 1.1 28.3 0.9 30.1 0.8 1.4 Arabuko Cynometra Forest 99 26.0 1.1 24.6 0.7 25.5 0.8 26.4 0.4 27.1 0.6 -0.7 Arabuko Human Disturbed 30.5 2.4 27.1 1.0 29.5 1.1 31.4 1.2 32.1 2.3 3.5 Arabuko Mixed Forest 99 26.4 1.2 24.9 0.7 25.7 0.9 26.9 0.3 27.6 0.7 -0.5 Ilcrct Bush 35.0 2.3 35.1 1.0 33.3 1.5 31.4 1.8 31.0 1.9 4.8 Ilcrct Grassland 34.4 2.2 35.1 0.5 34.7 1.9 33.8 2.1 33.9 1.7 4.1 Kakamcga Forest 94 18.7 0.5 18.2 0.3 18.3 0.4 18.8 0.5 19.2 0.4 0.2 Kakamcga Grassland 3 19.7 0.7 18.9 0.5 19.1 0.6 18.8 0.5 19.3 0.6 0.6 Mcru Bush 54 32.9 2.1 32.4 0.6 32.5 1.7 32.2 1.7 31.9 1.6 5.0 Mcru Human Disturbed 5 31.8 1.4 31.5 0.9 32.5 1.4 31.4 2.0 30.3 0.9 6.1 Mcru Forest 96 28.8 1.4 28.1 0.4 28.5 1.1 28.8 1.0 28.6 1.3 2.5 Mcru Grassland 0 30.7 0.8 30.7 0.8 31.4 1.6 30.3 2.0 29.2 0.7 2.7 Mt Kenya High Forest 10.9 0.4 10.7 0.5 11.1 0.3 10.8 0.4 10.5 0.4 0.3 Tabic 2.2 continued. T Site Canopy Cover (%) "AST oMAST ________________________________________ ( Q Mt Kenya Low Forest 100 12.7 0.7 Nairobi Bush 68 20.3 0.9 Nairobi Forest 86 19.5 0.9 Nairobi Grassland 0 22.6 1.6 Nakuru Forest 76.9 20.5 0.6 Nakuru Grassland 2.2 23.0 1.5 Shimba Hills Forest 94 23.5 1.0 Shimba Hills Grassland 0 28.6 2.5 Tana Forest 96 26.7 1.2 Tana Grassland 0 33.4 2.3 Tsavo East Bush 50 26.6 1.5 Tsavo East Forest 90 25.4 0.6 Tjja 0 tson tdjf JJA „ ^ SO N DJF 1 MAM \ V I ^ MAM 1 MAST 1 air V \ _______ \ __________________________________ 12.1 0.5 12.5 0.6 12.3 0.6 13.3 0.4 -23 17.9 0.7 19.4 0.9 20.2 0.8 20.5 0.6 0.7 17.6 0.6 18.6 1.0 19.6 0.9 19.4 0.6 0.6 19.8 0.7 21.3 1.2 11.3 0.9 11.3 0.9 3.0 20.0 0.5 20.1 0.3 20.6 0.6 20.7 0.6 0.3 21.9 1.0 22.0 0.9 23.0 1.3 22.9 1.3 2.8 22.5 0.6 23.2 0.8 24.7 0.4 24.7 0.5 -1.0 25.2 0.8 27.8 1.1 28.8 2.0 29.7 2.2 4.1 25.2 0.6 26.2 0.9 27.5 0.5 27.6 0.7 -3.2 30.0 1.1 31.8 1.4 34.7 1.1 33.6 1.6 3.4 24.5 0.5 26.4 1.5 27.4 1.0 26.8 0.7 -0.7 24.6 0.4 25.5 0.8 25.8 0.6 25.5 0.4 -1.8 to o Tabic 2.2 continued. Site Canopy Cover (%) T1 MAST CO MAST Tjja CC) ° J J A tson CQ SON TDJF CQ °DJF Tmam CC) °M AM T -T 1 MAST 1 air Tsavo East Grassland 0 29.9 1.7 27.8 0.7 29.5 1.4 30.5 1.8 30.3 1.1 2.7 Tsavo West Bush 34 26.7 1.6 25.5 0.7 28.1 1.3 26.4 1.2 25.4 0.8 23 Tsavo West Grassland 34 29.2 1.8 27.5 0.9 30.1 1.5 27.9 1.2 - - 4.8 Tabic 2.3 - Elevation corrected mean air temperature and precipitation measurements from nearby climate monitoring sites. Elevation was corrcctcd using an average environmental lapse rate of 6.8°C per kilometer. There is no correction for precipitation. Air Temperature "C Precipitation (mm) Site Elevation Correction JJA SON MAM DJF MAT JJA SON DJF MAM Total Arabuko Brachystcgia Forest -0.1 25.6 26.5 28.3 27.2 26.9 276.4 379.7 9.5 648.7 1314.3 Arabuko Cynomctra Forest -0.3 25.5 26.3 28.1 27.1 26.7 276.4 379.7 9.5 648.7 1314.3 Arabuko Human Disturbed 0.0 25.8 26.6 28.4 27.4 27.0 276.4 379.7 9.5 648.7 1314.3 Arabuko Mixed Forest -0.1 25.6 26.5 28.3 27.2 26.9 276.4 379.7 9.5 648.7 1314.3 Ileret Bush 0.5 30.5 31.3 30.0 29.2 30.2 0.0 21.0 112.7 134.2 267.9 Ileret Grassland 0.5 30.5 31.3 30.0 29.3 30.3 0.0 21.0 112.7 134.2 267.9 Kakamcga Forest -2.7 17.7 18.6 19.3 19.2 18.7 404.0 430.3 410.3 698.8 1943.4 Kakamcga Grassland -2.3 18.2 19.0 19.8 19.6 19.1 404.0 430.3 410.3 698.8 1943.4 Mcru Bush -1.4 26.5 27.8 28.7 28.9 28.0 6.7 117.9 138.3 280.0 542.9 Mcru Human Disturbed -3.5 24.4 25.7 26.6 26.7 25.8 6.7 117.9 138.3 280.0 542.9 Mcru Grassland -3.0 24.9 26.2 27.1 27.2 26.4 6.7 117.9 138.3 280.0 542.9 Mcru Forest -1.3 26.6 27.9 28.8 28.9 28.1 6.7 117.9 138.3 280.0 542.9 Mt. Kenya High Forest -7.3 10.6 11.0 11.2 10.9 10.9 97.6 145.6 195.9 357.9 797.0 to to Tabic 2.3 continued. Site Elevation Correction JJA Nairobi Bush 0.6 20.1 Nairobi Forest -0.1 19.5 Nairobi Grassland 0.7 20.2 Nakuru Forest 0.7 19.9 Nakuru Grassland 0.7 19.9 Shimba Hills Forest -2.4 23.3 Shimba Hills Grassland -2.4 23.4 Tana Forest 0.6 28.4 Tana Grassland 0.7 28.6 Tsavo East Bush 3.5 25.3 Tsavo East Forest 3.5 25.3 Tsavo East Grassland 3.5 25.3 Tsavo West Bush 0.7 22.6 Tsavo West Grassland 0.7 22.6 Air Tempcraturc°C SON MAM DJF MAT JJA Precipitation (mm) SON DJF MAM Total 20.1 17.9 20.1 19.6 303.5 724.0 118.3 186.7 1332.5 19.4 17.2 19.4 18.9 303.5 724.0 118.3 186.7 1332.5 20.1 17.9 20.2 19.6 303.5 724.0 118.3 186.7 1332.5 20.4 20.2 20.3 20.2 71.7 219.5 303.0 582.7 1176.9 20.4 20.2 20.3 20.2 71.7 219.5 303.0 582.7 1176.9 24.2 26.0 24.9 24.6 276.4 379.7 9.5 648.7 1314.3 24.2 26.0 24.9 24.6 276.4 379.7 9.5 648.7 1314.3 29.8 30.6 30.8 29.9 6.7 117.9 138.3 280.0 542.9 29.9 30.8 30.9 30.0 6.7 117.9 138.3 280.0 542.9 27.4 27.4 28.2 27.1 13.4 112.1 296.3 237.8 659.6 27.4 27.4 28.2 27.1 13.4 112.1 296.3 237.8 659.6 27.4 27.4 28.2 27.1 13.4 112.1 296.3 237.8 659.6 24.7 24.7 25.5 24.4 13.4 112.1 296.3 237.8 659.6 24.7 24.7 25.5 24.4 13.4 112.1 296.3 237.8 659.6 Because of the proximity to the equator for these sites, there is very little annual variation in mean daily soil temperature over the course of the year; the principal changes are due to the progression of the rainy seasons caused by migration of the ITCZ and the transition of the monsoons [Nicholson, 1996]. Throughout the soil temperature sites, the annual variance in deep soil temperature ranges from 0.5-2.5°C. The highest temperatures and variability within this data set are associated with open grassland and bushland locations where the mean annual average soil temperatures range from 20 to 35°C with an average standard deviation of 1.7° C. Conversely, the average soil temperature standard deviation in all the forest sites is 0.85° C. The hottest mean annual soil temperatures were recorded in the Turkana Basin near Ileret where daytime maximum temperatures at 0.25 m were >37 °C (Figure 2.2) during the warmest season. In the Turkana Basin mean annual soil temperatures were between 33 and 35°C (Figure 2.3). Seasonal averages of soil temperature at 25 cm depth and the associated seasonal standard deviations are presented in Figure 2.4. Seasonally there is much more variance in soil temperature, due mostly to fluctuations in soil moisture content caused by the wet and dry seasons and this varies greatly depending on the strength of the given season. Seasonally, the two hottest locations observed are within the Turkana Basin at Ileret and at the Tana River Primate Reserve. In the Turkana Basin, the hottest season averaged over the two years of data collection is during the months of June, July, and August. At the Tana River Primate Reserve the season with the hottest deep soil temperatures is observed during the months of December, January and February and the seasonal average during this time period is 33±1.2°C. 24 45 40 JJA SON DJF MAM June July Aug Sept Oct Nov Dec Jan Feb March April May 2009 2010 Figure 2.2 - One year of data from June 1, 2009 until May 20, 2010 at the Ileret Field station of the Turkana Basin Institute measured at 0.25 m depth. This figure illustrates the hottest deep soil temperatures observed during October 2009. to 40.0 35 o 30 03 CD c Qh 25 20 i it i U , c V * iU k r * * ,1 ti5t * 0* 15.0 10.0 lC-tO •o-< >" otr* §P •'C cn ctT (SS. ' o►n CD n&3 ►n o*C-D$ |-S CT O CD Q-►n o CD CD *C-D$ CD £PP• P£• S' 3- CD Cl* r2-, 0CQD 0CQD c/3 po po c8r ^ o O- CD $S3 CC//33 Q-CD CD a s Cd U M &S." 3CD* Q-CD S "Tl o c3 CD s $05■ CD £5 £L ffi I f - i-n o* CC-D/$3 £ £3 hTl O* C-D$ 2 •£-2t .' OO" ! $C=d 2S . 3 o►n 3 2 l£-2t .' o Q 2 R* c 3 ►n o»C-Dt 2 R* 3© §Q-Ct/ r2 C/2 >-] 3 Ef C/3 C/3 3 B* S3 &3 <1 <1 a- a- oCO o o &3 S3 O ffl ffl ffi ffi C•-D$ Cc/n3 CE/D3 CE/D3 C/3 C/3 § Cd Tfl O* C-D$ £D3 C/3 Q- $C/=3 =r O3 C/3 C/3 C/3 <O1 9£3 c>-n ] c/33 O O Cd Q 5 Pa §Q-Figure 2.3 - Plotted are the mean annual soil temperatures at 0.25 m depth with standard deviation plotted on the top of each bar. Colors correspond to type of vegetation cover where red indicates grassland localities, green indicates forested localities, and blue indicates bushland localities to On 40.0 27 h*-00« HHOHOl cm K JH Q ooo HHXaCH H O O +H i- C H C -O - h I-On OKD n-OOi 03 K )*tO H iKEDl KJOO txx> KXTH • hi-CH •OOCH HCtlQKH <mn .00 . m 10 CD CD O (<Nn (N0 < £ Q • 7- CIT, < Tsavo West Grassland Tsavo West Bush Tsavo East Grassland Tsavo East Forest Tsavo East Bush Tana Grassland Tana Forest Shimba Hills Grassland Shimba Hills Forest Nakuru Grassland Nakuru Forest Nairobi Grassland Nairobi Forest Nairobi Bush Mt Kenya Low Forest Mt Kenya High Forest Meru Grassland Meru Forest Meru Disturbed Meru Bush Kakamega Grassland Kakamega Forest Ileret Grassland Ileret Bush Mixed Forest Arobuko Disturbed Cynometra Forest Brachystegia Forest .0 <T> < too ^to cd to cd <u .2 C so ' 1 .2 £ .2> ^ cd G cd GO Coto cd V to o to uu tPo OO <N cd o CD O S .j ‘o a <u o cS CD <U <U Ictf °to tc§od J^ ^ J3 GO ^ i ^ (^n '>! £CD ^C3D 3 £O PM j o 9JHiBJ9dm9x red indicates SON, green indicates DJF, and purple defines MAM. Temperature and precipitation from surrounding weather stations are presented in Figure 2.5 and Figure 2.6, respectively. Climate information is not directly measured at the soil temperature sites and represents regional climate conditions. The difference in observed mean annual soil temperature (MAST) and mean annual air temperature (MAT) can be described by the soil temperature difference, which is simply the difference between the observed MAT and the MAST. The soil temperature difference is presented in Figure 2.7 and is described by Equation 1 28 \T _ T - T so il-a ir MAST MAT (1) where &Tsoil-air is the temperature anomaly Tsoil is the mean annual soil temperature and Tair is the mean annual air temperature that has been corrected for the elevation difference between the soil temperature sites and climate station location using the AELR. kTsoil-air is discussed further in section 2.4.1. Discussion Soil temperature relationship to ecology and local climate The spatial variation in the observed soil temperature can be broken down into two parts, as both local climate and ecology have important influences on soil temperature. In the study region, temperature variations are principally a function of elevation, as shown by a regression of the mean annual soil temperatures against altitude. It is necessary to separate the sites as the thermal regimes of forest soils are much different than those in bushland and grassland locations. 35 u 30 2S23 S<U - 25 £ 20 15 JJA SON DJF • MAM > §* %o GO O FT O FT CD CD ICD OQ S3 2 CD 2 £ 2 2 g o W . S 3 pg S CD ca | 5 1 - CS CD o ffl ES 05 * CO & * Figure 2.5 - Seasonal averaged 2 m air temperatures from nearby weather stations. The data were obtained from Kenya Meteorological Service. Colors are as in Figure 2.4. *NP=National Park, PR= Primate Reserve to VO Precipitation (mm) > ocCT" o in o ?r o CD Ci-Dl CD P P OCQD p CD $P 2 P _ o ' CT" 2 £ £ C/5 t r H p -5 £ a - p £ z g •-t C/3 $ <!O M p Figure 2.6 - Seasonal precipitation amounts at the weather stations near the measurement sites. Colors are as in Figure 2.4 O 6 8 Figure 2.7 - Seasonal temperature anomalies (mean annual soil temperature difference from mean annual air temperature). Colors correspond to the type of vegetation cover: Blue, bushland; green, forest; red, open grassland. Anthropogenically disturbed areas are shown in black. For the bushland and grassland the anomaly is almost exclusively positive indicating that the ground is warmer than the air. In forested localities the ground is cooler than air because of shading from the sun. The elevation and soil temperature relationships for Kenya are described by Equations 2 and 3 (Figure 2.8). Grassland/Bushland T (z) = 33.8 - 6.84z R2 = 0.75„ (2) Forest T(z) = 27.6 - 5.34z R2 = 0.94 (3). where z is elevation in kilometers. We now consider the relationship between canopy cover and soil temperatures. In addition to being a strong function of elevation, soil temperature should also be a strong function of vegetation cover because of the influence of direct solar radiation on the surface energy budget and thus, the amount of available energy for heating the soil surface. In open grassland settings there is little shade so there is direct exposure of the soil surface to solar radiation. In forested settings, most of the daily solar radiation is captured in the high canopy, thus shading the understory and the soil surface. In addition, the forest soils contain more water as the root zone acts to pull deep moisture to the surface through the process of hydraulic lift [Caldwell et al, 1998]. Surface shading and higher moisture content within a forest result in lower daily maximum soil and air temperatures. This effect is shown in Figure 2.9. Soil temperatures used in this figure were normalized to an elevation of zero using Equations 2 and 3 depending on the particular ecological setting. It is clear from this figure that canopy cover has a strong effect on the mean annual soil temperature, which can be described by Equation 4 Tsoil (x) = 34.4 - 0.07x . (4) where x is the percentage of forest canopy cover. As the percentage of canopy cover increases 32 33 Grassland Sites 40 A 30 20 O l0 V5 ­3 cd 5V-i ^ 0 ♦ ► ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ oo Forest Sites 3.5 Figure 2.8 - Mean annual soil temperature and elevation. A shows this relationship for Grassland and bushland vegetation types, B shows this relationship for Forest sites. 34 f-GO (N T3<D O (D-- o U T<3D 3 40 38 36 34 32 30 28 26 22 20 ♦ ♦ 10 20 30 40 50 60 70 Canopy Cover Percent 80 90 100 Figure 2.9 - The relationship of canopy cover to the annual mean soil temperature at 25 cm normalized to zero elevation. Temperatures were corrected using Equations 2 and 3 for the respective environments. 0 35 from opened to closed the mean annual soil temperature is greatly lowered. The effect of vegetation cover, particularly woody forest cover, can be explored further by referring to Figure 2.7 and looking at the soil temperature anomaly. Air temperatures in this figure were corrected for the elevation difference between the place of measurement and the soil temperature site. From this figure the effect of shading by vegetation cover is that soil temperatures are similar to or slightly less than air temperatures, whereas in open and bushland sites mean annual soil temperatures are much higher and have a larger deviation from air temperature. The formation of carbonates occurs when pore water percolating within the soil column becomes supersaturated with respect to calcite. For a full review of carbonate formation and techniques for extracting the temperature of formation, refer to Eiler [2011] and Cerling [1984]. The governing equation of the reaction is demonstrated in Equation 5 constant, and the brackets denote the activities of the aqueous species. It is assumed that water has an activity of 1.0. There are several ways in which carbonate formation is favored. First is a reduction in the partial pressure of carbon dioxide; this occurs during periods of high water stress on vegetation when plant growth is diminished and when root Soil temperature relationship to soil carbonate formation CaCO3 + P(CO2) + H2O = Ca2+ + 2HCO3 Kcal (5) where P(CO2) is the partial pressure of carbon dioxide in the soil, Kcai is the equilibrium respiration is decreased. Second is by increasing the activities of calcium or carbonate in the pore fluid; this results from chemical discrimination as water is removed from the system due to root uptake and when calcium and bicarbonate is excluded from uptake. Third, supersaturation can result from an increase in soil temperature, thereby reducing the solubility of carbon dioxide because of its retrograde solubility. Thus, these three drivers of carbonate formation all occur during periods of elevated soil temperature and aridity and enhance each other. It is therefore likely that the carbonate precipitation may have a strong seasonal bias toward the warmest and driest seasons [Breecker et al. 2009]. With conditions conducive to carbonate formation in mind, along with consideration of the soil temperature and precipitation data shown in Figures 2.5 and 2.6, respectively, the seasons in which conditions most likely to be representative of periods of carbonate formation start to become clear. Because carbonates are precipitated during the "drying out" phase of the soil column we expect that the temperature preserved during carbonate formation will be representative of hotter and dryer seasons. To better elucidate the temperatures present during the hottest and driest times of the year refer to Figure 2.10 where two years of the daily average temperatures observed at 25 cm depth in each microclimatological setting at four different study areas are plotted. In all cases the hottest temperatures are observed in the open setting. In areas such as Kakamega forest (upper right of Figure 2.10) the difference in deep soil temperature between the open and forest setting is small, ~1°C on average. In places such as the Tana River Primate reserve (lower right of Figure 2.10), the difference between soil temperature in the forest and in the open is ~7°C warmer than in the forest. 36 Jun Aug Oct Dec Feb Apr Jun Aug Oct Dec FebApr Jun Aug Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr 2010 2011 Figure 2.10- Daily averaged temperature at a depth of 0.25 m from the different microclimatological settings from four different national parks and reserves. Colors are as in Figure 2.3. It would be expected that soil carbonates would form, using the justification above, during the hottest parts of the year, which occur in the later part of the DJF season. Therefore, recording soil temperatures that are likely about 37-38 °C. However, in places such as Kakamega forest, it would be unlikely to observe much carbonate formation due to the large amounts of annual rainfall. Land use effects on soil temperature The areas surrounding the national parks in Kenya are prone to land use changes where pristine grassland and forests are removed for the purposes of agriculture. Over the past 50 years there has been a steady increase in the total land area used for agriculture such as cultivation and grazing, shown in Figure 2.11 [WorldBank, 2012]. The driving factor in this increase appears to be the steady increasingly rapid rise in human population in Kenya also shown in Figure 2.11. In this section we look at two parks that have experienced such land use changes. At Arabuko Sokoke National Forest the forest has been cleared for construction of a main road connecting Mombasa with Malindi, with small scale farming taking place between the road and the forest. At Meru National park, a small community lies along the park boundary where there has been a significant amount of cattle grazing, to the point of stripping the land to bare soil leaving only tall bushes. To assess the effects of these disturbances, soil temperatures were monitored in the disturbed areas. Referring to Figure 2.12 and comparing the community setting with the other ecological settings at Meru National Park, there are no easily discernable differences due to the effects of overgrazing. The differences observed at Meru are most likely due to the meteorological 38 conditions of the area during the study period. During that period there was a severe drought and limited vegetation coverage across the entire park. However, at Arabuko Sokoke National Forest (Figure 2.13) the effect of removing the forest is very clear. Comparison of each season reveals that the cleared areas are 2-5 degrees C higher than the surrounding forested settings. In conditions of the area during the study period there was a severe drought and limited vegetation coverage across the entire park. However, at Arabuko Sokoke National Forest, in addition to the higher seasonal averages, the seasonal standard deviation of the temperature is greater than is observed in the forest indicating that the cleared areas are much more prone to daily temperature fluctuations caused by more solar insolation reaching the soil surface. Conclusions Soil temperatures relevant to soil carbonate formation were evaluated for tropical soils in Kenya. Seasonal variation in soil temperatures is minimized by the small seasonal temperature fluctuations near the equator. Soil carbonates generally form at depths below approximatly 25 cm, by which depth the daily temperature fluctuation is attenuated to less than 2°C. Soil temperatures are related to site air temperature, with forested sites having values similar to air temperatures within approximatly 2°C, but bush and open sites have temperatures up to 7°C higher than average meteorological air temperatures. These results confirm that high soil temperatures (30 to 40°C) recorded by paleosol carbonates in the Turkana Basin using the A47 method [Passey et al., 2010] are in the range of observed soil temperatures for hot, arid regions of East Africa. Land use change results in higher soil temperatures, especially where forests are cleared for 39 40 farming. Soil temperatures in disturbed, previously forested areas were approximatly 4°C hotter than the equivalent ecosystem in the Arabuko Sokoke Forest. Land Area (x 1000 sq km) 41 - Agricultural Area - Population 275 45 1960 1970 1980 1990 2000 2010 Figure 2.11 - Total land area in Kenya that has been designated for agriculture and population during the past 60 years. Data comes from the World Bank (2012). Population (Millions) Temperature °C 42 40.0 35.0 30.0 25.0 20.0 OJJA OSON ODJF OMAM Meru Bush Meru Disturbed Meru Forest Meru Grassland Figure 2.12 - Seasonal averaged soil temperatures at all four Meru vegetation settings. The community setting is very similar to the bush site as there is little grass and it is characterized by predominantly acacia bushes. The effects of grazing are not prevalent between the grassland and bushland sites in this plot, however, this sampling interval was skewed by a significant drought in the area. Temperature (°C) 43 Figure 2.13 - Mean annual temperatures at Arabuko Sokoke National Forest. This figure illustrates the changes in the thermal regime of the soil because of deforestation 44 References Breecker, D. O., Z. D. Sharp, and L.D. Mcfadden (2009), Seasonal bias in the formation and stable isotopic composition of pedogenic carbonate in modern soils from central New Mexico, USA, Geological Society of America Bulletin 121(3-4): 630-640. Caldwell, M. M., T. E. Dawson, and J.H. Richards (1998), Hydraulic lift: Consequences of water efflux from the roots of plants, Oecologia 113(2): 151-161. Cerling, T. E. (1984), The stable isotopic composition of modern soil carbonate and its relationship to climate, Earth and Planetary Science Letters 71(2): 229-240. Cerling, T. E., N. E. Levin, and B.H. Passey (2011), Stable isotope ecology in the Omo- Turkana Basin, Evolutionary Anthropology 20(6): 228-237. Cerling, T. E., N. E. Levin, J. Quade, J.G. Wynn, D.L. Fox, J.D. Kingston, R.G. Klein, F.H. Brown (2010), Comment on the paleoenvironment of Ardipithecus ramidus, Science 328(5982): 1105. Kenya Meteorological Department (2012), Climatological Data, 2012, from http://www.meteo.go.ke/data/. Eiler, J. M. (2007), "Clumped-isotope" geochemistry-The study of naturally-occurring, multiply-substituted isotopologues, Earth and Planetary Science Letters 262(3-4): 309-327. Eiler, J. M. (2011), Paleoclimate reconstruction using carbonate clumped isotope thermometry, Quaternary Science Reviews 30(25-26): 3575-3588. Geiger, R., R. H. Aron (2009), The Climate Near the Ground, Rowman & Littlefield, Lanham, MD. Ghosh, P., J. Adkins, H. Affek, B. Balta, W. Guo, E.A. Schauble, D. Schrag, J.M Eiler (2006), 13C-18O bonds in carbonate minerals: A new kind of paleothermometer, Geochimica et Cosmochimica Acta 70(6): 1439-1456. Griffiths, J. F. (1972), Climates o f Africa, Elsevier Pub. Co., Amsterdam, New York. Jury, W. A. and R. Horton (2004), Soil Physics, John Wiley, New York. Lal, R. and M. Shukla (2004), Principles of Soil Physics, M. Dekker, New York. 45 Nicholson, S. E. (1996), A review of climate dynamics and climate variability in Eastern Africa. The Limnology, Climatology and Paleoclimatology of the East African Lakes, Gordon and Breach: 25-56, Amsterdam, The Netherlands. Passey, B. H., N. E. Levin, T.E. Cerling, F.H. Brown, and J.M. Eiler (2009), High-temperature environments of human evolution in East Africa based on bond ordering in paleosol carbonates, Proceedings of the National Academy o f Sciences. 107(25): 11245-11249. Ratnam, J., W. J. Bond, R. J. Fensham, W. A. Hoffman, S. Archibald, C. E. R. Lehmann, M. T. Anderson, S. I. Higgins, and M. Sankaran (2011), When is a ‘forest' a savanna, and why does it matter?, Global Ecology and Biogeography 20(5): 653-660. Wallace, J. M. and P. V. Hobbs (2006), Atmospheric Science: An Introductory Survey, Elsevier Academic Press, New York. White, F. (1983), The Vegetation of Africa, Natural Resources Research, Vol. 20 United Nations Scientific and Cultural Organization, Paris. White, T. D., S. H. Ambrose, G. Suwa, and G. Woldgabriel (2010), Response to comment on the paleoenvironment of Ardipithecus ramidus, Science 328(5982): 1105. World Bank, Agricultural Land (sq. km), Food and Agriculture Organization, electronic files and web sites, http://data.worldbank.org/, retreived from web 2012. World Bank, Population Total, electronic files and web sites, http://data.worldbank.org/, retreived from web 2012. CHAPTER 3 TROPICAL ENVIRONMENTAL SURFACE TEMPERATURES PRESENTED AS A COMPOSITE DAY Abstract The soil temperature across any landscape can vary greatly and have a large influence on the daily routines of many of the floral and faunal inhabitants. In the following study we compare soil surface temperatures from three different microclimates that are common in Kenyan landscape mosaics: grasslands, bushland, and forest. Soil surface temperatures were calculated from subsurface soil temperature profiles using an iterative inversion solution to the 1 dimentional heat diffusion equation. The results are presented as a statistical composite day, which effectively shows the average evolution of an equatorial diurnal cycle. The temperature difference observed between surface temperatures present in open grassland settings and those present in forests is often on the order of 20°C during the hottest parts of the day. During the night the temperature difference between the different microclimates is reduced to near zero. In the open the average daily temperature range is 25-30°C, whereas in the forest this range is 5-10°C Introduction It has been proposed that hot settings, such as the Turkana Basin in Northern Kenya, the Awash Valley in Ethiopia, and Olduvai Gorge in Northern Tanzania were main stages for the development and evolution of early hominids [Bobe, 2009]. Preliminary paleothermometry results from East African hominid fossil sites using the A47 method suggest that hot environments with average soil temperatures between 30 and 40°C have existed in these areas for much of the past 1-4 million years [Passey, 2009] and today, environments located within the Turkana Basin are amongst the hottest 1% of places worldwide with a mean annual air temperature of 29.2°C [Hijmans, 2005]. If hominids did evolve under such extreme temperature conditions, they would have had to deal with considerable thermal stress. Thermoregulatory advantages of modern humans such as sweating, bipedalism, and reduced body hair are thought to have resulted from evolutionary adaptations to exposure in such extreme hot and dry environments [Ruxton and Wilkinson, 2011; Wheeler 1984, 1991, 1992; Ruff, 1991). In addition to these proposed adaptations there is evidence to suggest adaptations of the human brain [Falk, 1990] and skin [Jablonski and Chaplin, 2000, 2010] also aided in dealing with the high thermal heat load that early hominids would have been regularly exposed to in tropical midday environments. Although significant, these adaptations alone are not sufficient to have enabled hominids to survive in hyperthermic environments such as those found in sub-Saharan Africa today. It would have been likely, as it still is for modern humans today to take advantage of the spatial thermal differences and take refuge in cooler shaded areas such as a forested riparian zone or within the shade of local flora. 47 Across landscapes temperatures within different microclimates can be very different and are a function of the amount of shade provided by the surrounding flora [Geiger, 1965]. Although the differences between the different microclimates are not documented well, human and animal behaviors suggest that they can be quite large, as many seek shade during the peak of the diurnal temperature cycle. The landscape in which hominids resided is debated, and ranges from closed forests to open savannas [Cerling et al, 2010; Potts, 1998; White et al. 2009; White et al, 2010]. Most evidence in the sedimentary record in these areas suggests that the predominant settings would have been riparian zones, deltaic environments, floodplains, and lake margins [Feibel, 2012; Feibel et al, 1991]. In the following chapter we describe continuous subsurface soil temperature profiles collected over a two year period from June 2009 until May 2011 from which we estimate the soil surface temperature and illustrate the strong environmental differences present. The areas selected for this study were chosen because they are thought to be close representatives of the types of environmental composites present in the Turkana Basin, Awash Valley, and Olduvai Gorge 1-4 million years ago during the time when humans evolved. We use soil temperature profiles from open, bush covered, and forested sites in two lowland sites, Meru National Park and Tana River Primate Reserve, and one highland site, Nairobi National Park. Using soil diffusivity models, we invert the soil temperature time series to derive the soil surface temperature time series. We use the acquired surface temperature time series to produce a "statistical composite day" whereby, seasonal hourly averages of the ground surface temperature represent a seasonally 48 averaged day within different microclimates. We then discuss these results in the context of environmental differences at the landscape scale. Methods and Materials Soil temperature data collection Soil temperature measurements were collected using HOBO Pendent® data loggers; these have a stated measurement range of -20°C - +70°C, with an accuracy of ±0.5°C and a precision of 0.1°C. At each site the temperature loggers were buried in a 0.1 m wide trench at three depths: 0.05, 0.15, and 0.25 m. After the sensors were placed the trench was then back filled with the same soil material that was removed. Where possible, three distinct environmental types, forest, grassland, and bushland environments, as described by the UNESCO classification scheme [White, 1973], were targeted. In some cases, disturbed areas related to land use change were also monitored. The loggers were programmed to record an instantaneous temperature measurement once every 20 min. The soils were monitored between May 2009 and May 2011. Site selection and description Latitude, longitude, and elevations for the study sites are given in Table 3.1 and a DEM with the study sites is presented in Figure 3.1. Meru National Park (MNP) is located on the equator in Central Kenya on the eastern flank of the East African Highlands. The average elevation of the park is about 400 m and covers an area of ~900 km2 that is composed of mostly open grassland and bushland. The Tana River runs along the southern edge of Meru Park along the boundary 49 with Mwingi National Reserve and Kora National Park to the south. Along the river there is a dense riparian forest that varies in width from a few meters to tens of meters. Within this riparian forest the average canopy coverage is about 70% as calculated from circular fisheye photography. Soil temperatures were logged at three locations within this park, one in grassland, one in bushland, and one in a riparian forest zone. Nairobi National Park (NNP) is located at approximately 1.3° S latitude. It is situated south of the city of Nairobi at an average elevation of 1650 m. The park covers an area of about 117 km2 and is composed of mostly grassland to the southeast and forest to the northwest with a transition from grassland to bushland to forest over several 100 m. A dry deciduous forest is situated in the extreme northwestern part of the park with average canopy coverage of 86%. As at Meru, instruments were placed in open grassland, under heavy bush cover, and in a forested site. The Tana River Primate Reserve (TRPR) is located at approximately 1.8° S latitude in Eastern Kenya along the Tana River. The reserve includes a dense, closed-canopy riparian forest along the Tana River with marginal grasslands and bushland. The riparian zone in this area ranges in width from a few tens of meters to several 100 m. Canopy coverage throughout the forest zone is ca 90%. This riparian forest is home to two endangered endemic species of monkey, the Tana River Mangabey (Cercobus galeritus) and the Tana River Red Colobus Monkey (Procolobus rufomitratus). At this location, records were taken at two sites, one in open grassland/bushland and the other within the dense riparian forest. 50 Table 3 .1 - Latitude, longitude, and elevation o f the selected soil temperature sites. Site Latitude Longitude Elevation (m) Mem Bush -0.07013 38.41288 349 Meru Open 0.18012 38.22673 596 Mem Forest -0.07175 38.42017 336 Nairobi Bush -1.35137 36.79636 1701 Nairobi Forest -1.34836 36.76731 1807 Nairobi Open -1.35143 36.79628 1695 Tana Forest -1.87652 40.13994 42 Tana Open -1.87615 40.13791 37 52 ' I__JL_I__I__ Low : 0 Kilometers Figure 3.1 - Digital elevation model showing the locations of the study sites. Description of the seasons The data from Meru National Park and TRPR were collected for one year from May 2009 until May 2010; those from Nairobi National Park were collected the next year, from May 2010 until May 2011. From these data we present calculated land surface temperatures for different environments within these Kenyan national parks. The data are presented as yearly and seasonal averages of each hour of the day creating what we refer to as a statistical composite day. Seasons are defined by the Kenya Meteorological Service as: 1) a "Warm Dry Season" (December, January and February; (DJF)) where very little rainfall is observed; 2) the "Long Rains" (March, April, and May (MAM)) during which a majority of Kenya's annual rainfall occurs; 3) a Cool, Dry Season (June July and August (JJA)); and 4) the "Short Rains" (September, October and November (SON) which is a secondary rainy season associated with the migration of the Inter- Tropical Convergence Zone (ITCZ) over the area [Griffiths, 1972; Nicholson, 1996]. Further details of the meteorology of Kenya are given in Chapter 1. Data Analysis Procedure Here we employ a mathematical inversion method in order to obtain an accurate estimation of the soil surface temperature from continuous subsurface measurements. In summary, we use the theory of soil heat flow to calculate the thermal diffusivity of the soil [Jury, 2004] then we follow the methods described by [Barttlet et al., 2006] to iteratively solve for the surface temperature input function from a fixed depth of 5 cm. Temperature propagation into the soil column is described by the 1 dimensional heat diffusion equation, 53 d d2 - T (z, t) = a eff - T (z, t) (1) dt dz where T is the temperature, t is the time, z is depth, a f s the effective thermal diffusivity [Carslaw and Jaeger, 1986]. Since a temperature time series can be represented as a series of step changes we can employ the error function solution shown of Equation 1 described by Carslaw and Jaeger [1986] 54 i T(z,t) - 2 (T„ - T e r f c (2) _^ 4 * a eff ( t n - t 0 ) where Zobs is the depth of observation and erfc is the complementary error function. The value of aef was not directly measured, therefore it was necessary to calculate aeff using a regression of the average daily amplitude against depth. The daily amplitude of soil temperature decreases exponentially with depth [Jury, 2004] following Equation 3. A(z) = A0e~kz (3) where A(z) is the amplitude of the daily temperature signal as a function of depth, z is depth; Ao is the amplitude of the daily temperature signal at the surface. aejf is a material property of the soil and describes how well a heat pulse can propagate through it. The value for k in Equation 3 is the wave number of the thermal wave and is a function of aeff and the period of the wave Equation 4. a * T Equation 4 can be transformed into slope intercept form by taking the natural logarithm of the right and left and sides of Equation 4, yielding Equation 5, ln( A( z)) = -kz + ln( A0) (5). Equation 5 suggests that we can take the amplitudes at three different depths and perform a linear regression on the natural logarithms of those amplitudes as a function of depth. The resulting slope of this regression line is a function of aeff, which can be obtained through Equation 4. The diffusivity values used in the calculations of surface temperature are listed in Table 3.2. Now that the value of the soil thermal diffusivity is known, we use Equation 2 to calculate the surface temperature input function following the methods described in [Bartlett et al., 2006]. The drawback in using Equation 2 for such modeling is that it is exclusively a forward solution; that is, it only works to diffuse temperature down into the soil column and forward in time. In order to solve for the surface temperature we need to calculate a higher, shallower level in the soil column and, because of the phase lag, backwards in time. To address this problem it is possible to use an iterative method with Equation 2 using an initial guess for the surface input then diffusing it into the ground, using Equation 2 and comparing the fit with the observed time series. The initial guess for the surface temperature was obtained by diffusing the 0.05 m observed temperature down to 0.1 m using Equation 2. The difference in the amplitude decay and time shift of the thermal wave between the observed temperature and the diffused temperature were used as scaling factors to scale the observed 0.05 m soil temperature time series to the surface as the initial guess. The initial guess is then diffused into the soil column, using Equation 2, to the observation depth of 0.05 m. The misfit between the two time series is then used to adjust the input function for the next iteration and the process is repeated. Typically the convergence on a local minimum in the residuals was observed within 5-10 55 iterations of the soil temperature model (Figure 3.2). Once the minimum in the residuals is reached the result for the surface temperature has been reached. The code for this procedure is written in Matlab. Model output for a one-month period at the TRPR grassland/bushland site predicts surface temperatures well for most of the time series; Figure 3.3 shows that the difference between the 0.05 m time series modeled from the surface input and the actual observed 0.05-m time series is commonly less than 1°C. From the figure it is observed that the largest errors occur regularly at the daily maximum. During events that cause large deviations in soil temperature (Fgure 3.3, panel B) such as heavy rain, the model performs less well as the error during these periods is larger. During periods of rain the thermal diffusivity of the soil is temporarily disturbed because of the additional water infiltrating into the soil column and the dependence for diffusivity on soil moisture content. In addition to an alteration of the thermal diffusivity, water flux into the soil column acts to advect soil heat with it [Jury, 2006]. Composite Day We define a "composite day" as the average hourly temperature over a 24-hour period for each of four 3-month seasonal blocks. In this study, these time intervals correspond to DJF, MAM, JJA, and SON . We take the individual daily surface ground temperatures, as determined using the inversion methods described above, to obtain hourly temperatures of the ground surface. 56 57 Iteration Figure 3.2 - RMS Error as a function of iteration number of the diffused soil temperature signal compared with the observed soil temperature signal at 5 cm depth. The minimum of the RMS error is at iteration number 7. The high value of 50 that is observed as the first point is an arbitrary number used to initialize the model. 58 Table 3.2 - Values for diffusivity (otcfr) for selected soil temperature sites Site Diffusivity (m*m/s) Uncertainty (±) Nairobi Grasslan 6.4E-7 ±2.0E-7 Nairobi Bush 1.8E-7 ±5.4E-8 Nairobi Forest 3.8E-7 ±1.1 E-7 Tana River Grassland 2.6E-7 ±8E-8 Tana River Forest 2.9E-7 ±9E-8 Meru Grassland 1.3E-7 -L4E-7 Meru Bush 4.7E-7 ±1.4E-7 Meru Forest 3.7E-7 ±1.1 E-8 59 Residuals Figure 3.3 - One month of modeled data from the Tana River Primate Reserve open site during May 2009. A) Misfit between the modeled 5 cm temperature, using the calculated surface temperature as the input, and the observed temperature time series at 5 cm depth. B) Modeled and observed temperatures at 5 cm. Results Composite day statistics provide good descriptions of what an average day looks like throughout the course of a year at different sites, with notable variations due to local ecology. All temperatures discussed in the ensuing text in this section are soil surface temperatures. The daily minimum temperature is observed just before sunrise at about 0600 local time throughout the year. At the time of the diurnal minimum, surface temperatures in all environments (forest, bush, grassland) are close to being in radiative equilibrium with the atmosphere, so these temperatures are very similar with differences among the sites due to differences in the atmospheric water vapor burden. After sunrise, the surface warms gradually throughout the day as absorbed solar energy heats the soil surface until it reaches a peak at about 1400 local time. As the local afternoon progresses, the net radiation becomes negative and the surface begins to cool. During the night, the open environments cool slightly more than the forested sites and bush sites because the surface of the open sites is able to radiate freely to the atmosphere, whereas coverage from forests and bushes thermally insulates the soil and prevents radiative heat loss of the surface. Cooling occurs gradually at first and then rapidly as the solar angle gets smaller and the sun eventually crosses the horizon. After sunset, which occurs at about 1800 local time, the surface continues to cool and temperatures within the different environments converge as they approach radiative equilibrium with the atmosphere. The full time series for the grassland, bush, and forest site at Meru National Park and the seasonal composite days for the four seasons allow comparisons of the composite day with the overall trend (day to day occurrences) of the season (Figure 3.4). The first few months, JJA, of the observation period were exceptionally hot and dry due to the 60 drought of 2009. During these months, the open and bush sites experienced large diurnal ranges and extremely high daily maximum temperatures because the soils within these areas were dry and much of the incident solar energy directly heated the soil column. The forested site during this period was much cooler throughout the day despite the drought, and it had a much lower daily maximum temperature and a smaller diurnal range. All of these features are evident in the composite day for the JJA season within the three environmental types. The remaining seasonal composites compare similarly with the first season, JJA. The exception to this, however, is the last season (MAM), where there is a distinct difference in the daily amplitude throughout all the environments as it is smaller and much more comparable than in previous seasons. This season is characterized by cooler and wetter climatic conditions in Kenya so this behavior is not a surprise. However, the temperature climatology during these cooler and wetter periods suggest that it may be possible for species that are normally confined to forest environments (because of daytime heat and nighttime predation) are able to travel across open areas for extended periods of time without the consequences that would exist during the warmer seasons. The results from the TPRP reveal temperatures that are equally as hot as were observed at Meru and, as well, exhibit a sharp temperature contrast between forested and open environments that is obtained throughout the year (Figure 3.5). The hottest season recorded at Tana River was during December, January and February, which is consistent with the seasonal definitions discussed in the introduction to this chapter. During this season, the temperatures in the open came close to or exceeded 60°C (140°F) regularly. The composite day for this locality shows that between the hours of 1000 and 1800 temperature in the open usually exceeds 40°C (>100°F) with an average daytime 61 maximum during the hottest season exceeding 50°C. The average diurnal temperature range during this period is ~20-25°C in the open. Within the forest zone, despite the extreme conditions in the open areas, the average diurnal range is approximatly 5°C throughout the year and temperatures only reach daily maxima of about 30°C. This is reflected in the composite day for this environment, as the average daily high during all four seasons is less than 30°C. If the open environment is compared to the forested environment this is a 20-25°C temperature difference, on average, between open and forested environments during the hottest parts of the day. The third park included in this discussion is Nairobi National Park. Nairobi National park is located in Central Kenya and is over 1000 m higher than the previous two sites just discussed. Therefore, it can be expected that, because of adiabatic cooling of the atmosphere, temperatures will be lower throughout the year than temperatures observed in Meru National Park or in the Tana River Primate Reserve (Figure 3.6). Despite lower average temperatures, the thermal difference between different environments is not as distinct at higher altitudes. At Meru and Tana River, where the temperature differences between the forest and the open were 20°C or more, the temperature contrast at Nairobi National Park between the environments is only about 5- 10°C during the hottest season, DJF, and less than 5°C during the coolest months. Daily high temperatures in the open there rarely get above 40°C, whereas at the other locations the temperature in similar settings soared over 60°C. 62 Cold Dry Season JJA ihort Rains SON June "09 Aug ‘09 Oct *09 D e c *09 Warm Dry Season Long Rains DJF MAM -Open -Bush Fo rest Feb * 10_____________May' 10 60 40 20 0 2 4 6 8 10 12 14 16 18 20 22 0 60 E MAM ‘10 0 2 4 6 8 10 12 14 16 18 20 22 0 0 2 4 6 8 10 12 14 16 18 20 22 0 Ilour Ilour Figure 3.4 - A) One year of modeled surface temperature for the different environmental types at Meru National Park. B-E) Seasonal composite days for the different environmental types at Meru National Park. Colors correspond to type of vegetation cover where red defines grassland localities, green defines forested localities, and blue defines bushland localities. P 60 <L>- c3 40 uoCu - ■2 20 Grassland Bush Forest JJA 409 G\ Cool Dry Season Short Rains Warm Drv Season Long Rains JJA ------------- DJF - Open Forest June ‘09 Aug ‘09 Oct ‘09 Dec ‘09 Feb ‘ 10 May' 10 0 60 o> L-3 40 s . 5201 Grassland Forest B JJA‘09 0 2 4 6 8 10 12 14 16 18 20 22 0 0 2 4 6 8 10 12 14 16 18 20 22 0 40 20 ; e MAM ‘10 0 2 4 6 8 10 12 14 16 18 2 0 22 0 Hour Figure 3.5 - A) One year of modeled surface temperature for the different environmental types at Tana River Primate Reserve (TRPR). B-E) Seasonal composite days for the different environmental types at TRPR. Colors correspond to type of vegetation cover where red defines grassland localities and green defines forested localities. DJF‘10 2 4 6 8 10 12 14 16 18 20 22 0 Hour O n A B- 60 40 Grassland Bush Forest JJA ‘10 B 60 40 SON -10 C Figure 3.6 - A) One year of modeled surface temperature for the different environmental types at Nairobi National Park. B-E) Seasonal composite days for the different environmental types at Nairobi National Park. Colors correspond to type of vegetation cover where red defines grassland localities, green defines forested localities, and blue defines bushland localities. Cool Dry Season Short Rains Warm Dry Season Long Rains Ju n e ‘ 10 Aug M0 O c t ‘10 Dec ‘ 10 Feb ‘ 11 MAM1 -Open -Bush Forest May' 11 Discussion Validation of high soil surface temperatures In an effort to validate the high calculated surface temperatures, a series of smaller experiments were conducted wherein a temperature sensor was placed directly on the soil surface. This method should provide a reasonable measure of the temperature at the ground surface. Because the HOBO temperature sensors are plastic and colored similarly to the surface, placing one directly on the surface should allow it to obtain the same, or at least a very similar temperature. This is shown in the following figure, where two sensors were placed close to one another for one day, one being on the surface and one being covered by 1-2 mm of soil. The temperatures of both agree well throughout the day in the sunlight and only differ during the presence of clouds, which are indicated through changes in the light intensity also plotted on Figure 3.7. Surface temperatures, based on sensors lying on the ground surface were measured for one month in the Turkana Basin; the results are shown in Figure 3.8. From this figure it can be seen that during 21 out of 23 of the days during this period, surface temperatures exceeded 50°C; furthermore, the surface temperature on 8 of those days exceeded 60°C. To further elucidate the difference that the shade provided by vegetation cover can have on the soil temperature, sensors placed under the shade of a nearby acacia tree and under a Salvadora bush during the same period of time showed a significant decrease in observed surface temperature. During the same time period, where daily maximum temperatures in the open exceeded 50°C, there were no days in the shade of either the acacia tree or the Salvadora bush where the daily maximum temperature was greater than 50°C. On most days, the surface temperatures were 15-20°C cooler in shaded areas than they were in the 66 open. Air temperature data from a weather station about 100 m away show that differences between the soil surfaces in different vegetative environments can be very large, especially in open exposed areas. In the open, the average temperature difference from the soil surface to 2 m above the surface is 24.3°C with a standard deviation over the period of 3.7°C. In the shade however, this difference is less by about a factor of 2, where the average difference over the observation period is ~9°C under the acacia tree and ~5.5°C under the Salvadora bush. The difference between the soil surface temperature and air temperature at the sites presented in the Results section is discussed further in the next section. Air temperature and soil temperature comparison Although the results presented above deal exclusively with soil temperatures, soil temperature and air temperature differ from each other quite significantly during the daylight hours and are comparable during the night. In addition, soil temperature data are difficult to obtain unless a study is conducted that deals specifically with soil temperatures to some extent. It is rare to find climate monitoring stations that operate soil temperature probes. What is more widely available are climate data such as air temperature from surface weather stations. The following discussion assesses how well these two values, soil and air temperature compare. Daily maximum and minimum air temperatures were collected from climate monitoring stations near Nairobi National Park and TRPR. The mean annual and seasonal averaged daily maximum and minimum temperatures for the soil surface and air are presented in Table 3.3. 67 Temperature °C 68 Figure 3.7 - One day of data comparing the accuracy of a soil temperature sensor that is exposed directly to the sun versus a sensor that is covered by 1-2 mm of soil. Light intensity is also plotted to show the presence of clouds. Light Intesity (lux) X 1000 69 Figure 3.8 - Measured daily maximum surface temperatures at TBI Turkwel station from October 31 until November 21 2011. Data is shown for an open site, a site in the shade of an acacia tree, and a site in the shade of a Salvadora bush. Also plotted in this figure are the daily maximum air temperatures from a weather station about 100 m away. The daily maximum soil surface temperature at Nairobi differed from the measured air temperatures to different degrees depending on the type of vegetation cover. In the open, the ground surface generally reaches maxima 10-15°C hotter than are measured in the air. In the forest, the temperature differences between the soil surface and air temperature are smaller (~2-5°C) as the forest surface is shaded from the sun and stays cooler. As expected, the daily minimum temperatures are quite comparable between all of the measured values. The bush site and the forested site stay warmer than the observed air temperature. This is most likely due to the vegetation cover in both of these settings acting to thermally insulate the ground surface. In the open, the effect is opposite as the lack of vegetation cover allows for rapid radiant heat loss to the atmosphere. At Tana River, the difference in the observed air temperatures and the ground surface temperatures follow the same pattern as at Nairobi; the maximum daily temperatures are quite different and the minimum daily temperatures are similar. At Tana River however, the closest climate station was 150 km away and about 100 m lower in elevation; therefore, the air temperatures presented in Table 3.3 have been corrected for this difference in altitude using an average environmental lapse rate of -6.5°C/km [Wallace and Hobbs, 2006] . In the open, the differences in the average maximum temperatures between the soil surface and the air were ~20°C, whereas the average maximum surface temperatures of the forest floor is cooler than the maximum air temperatures by ~4-5°C. Climate data that are available from the KMO is only daily maximum and minimum for the sites considered [Kenya Meteorological Service, 2011]. However, from the experiment that was conducted at TBI Turkwel and is discussed later, we can create 70 Table 3.3 - Annual and seasonal mean dally maximum, minimum, and ranges in both soil surface and air temperatures for Nairobi National Park, Tana River Primate Reserve, and Meru National Park. Air temperatures were collected from a climate monitoring station in close proximity to the soil temperature sites. The air temperatures have been corrected for the elevation difference between the climate monitoring station and the soil temperature sites using the average environmental lapse rate of 6.5'C/km. At Meru, the air temperature has been corrected to the elevation of the open soil temperature site, which is situated about 300 m higher than the bushland and forested sites. Nairobi Air Nairobi Soil Open Nairobi Soil Bush Nairobi Soil Forest Tana Air Tana Soil Open Tana Soil Forest Meru Air Meru Soil Open Meru Soil Bush Meru Soil Forest Annual Max 25.1 37.8 31.5 27.3 33.9 54.2 29.8 39.2 48.3 41.1 30.8 Min 14.4 15.3 15.8 16.2 23.1 23.6 24.5 28.5 23.4 26.6 25.4 Range 10.7 22.5 15.6 11.0 9.9 30.6 5.3 10.8 24.9 14.4 5.4 JJA Max 22.9 34.1 27.0 24.1 32.3 49.7 28.2 37.6 56.0 44.6 32.1 Min 13.6 14.5 14.9 15.3 21.8 22.0 23.0 27.1 22.4 26.6 25.5 Range 9.3 19.6 12.1 8.7 9.6 27.6 5.2 15.0 33.6 18.0 6.6 SON Max 25.3 39.6 34.6 27.5 33.9 53.9 29.7 39.2 51.3 44.1 31.2 Tabic 3.3 continued. Nairobi Air Nairobi Soil Nairobi Soil Nairobi Soil Tana Open Bush Forest Air Min 14.3 14.6 15.5 16.0 22.8 Range 11.0 25.0 19.1 11.4 10.2 DJF Max 26.8 41.5 35.5 31.9 35.0 Min 14.6 15.6 16.0 16.6 23.7 Range 12.1 25.8 19.4 15.3 10.5 MAM Max 26.5 37.1 30.0 26.5 34.4 Min 15.7 16.8 17.3 17.1 24.1 Range 10.8 20.2 12.6 9.3 9.4 Tana Tana Meru Meru Meru Mcru SoU SoU A.r SoU SoU SoU Open Forest Open Bush Forest 22.5 31.4 59.6 26.4 33.3 53.1 23.6 29.5 23.9 5.8 31.3 25.5 5.8 29.8 25.5 4.3 28.2 15.5 40.4 29.1 15.8 39.7 29.4 14.8 23.6 27.7 47.4 24.2 23.3 37.8 23.1 14.7 26.7 17.4 40.9 27.5 13.5 34.2 25.4 8.7 25.3 5.9 31.6 25.7 5.9 28.2 25.1 3.1 to composite days of both soil temperatures and air temperatures over the same period to assess the difference from the soil surface to 2 m throughout the day (Figure 3.9). We chose 2 m for two reasons: 1) because this is the height of the air temperature sensors and is specified by the World Meteorological Organization; 2) because this is close to the average height of most humans and thus, their entire bodies would be exposed to this temperature difference. From the figure it can be seen that, in the open, the maximum difference between the surface and air temperature is ~20°C, whereas in the shade it is smaller with a maximum difference of -5°. Soil temperatures deviate from the air temperature throughout the daylight hours and they reach their greatest difference when the sun is at or just beyond its' daily zenith. Air temperature lags soil temperature slightly because the air temperature is forced by the temperature of the soil. Under the dense shade of the Salvadora bush, soil surface temperature and air temperature are similar throughout the day. They do, however, deviate slightly in the afternoon and evening, which is likely due to a change in the sun angle and a change in the amount of light penetration through the bush. Relationship to Human Heat Tolerance Since the last common ancestor (LCA) between humans and other primates, a series of thermoregulatory adaptations have given hominids and modern humans the ability to venture into extremely hot environments without becoming dangerously overheated. Hominids and humans exhibit the continuation of an evolutionary trend toward higher heat loss capacity due to sweating and the evaporative cooling of the body provided by this mechanism [Eichna, 1950]. In order to maintain acceptable core body 73 Temperature °C 74 Hour Figure 3.9 - Comparison of composite days composed from the test period October 30th through November 21st 2011 at TBI Turkwel Station. Soil surface temperatures were collected by placing a sensor directly on the surface, and air temperature was collected over the same period from a weather station about 100 m away. temperature of 36.6-37.3°C effectively without a thermoregulatory response, the ambient environmental temperature must be between 24-29°C [Kerslake, 1972]. Within this temperature range a bare skinned individual, at rest, will be in a state of thermal equilibrium with its surroundings. Any activity or exposure to higher temperatures results in a response of the thermoregulatory system (i.e., sweating will begin). Exposure to temperatures greater than 40°C, could begin to overload the ability for the body to cool itself, resulting in a dangerous rise in core temperature [Kerslake, 1972]. Analysis of the composite day for different environmental types show the temperature conditions that require only a minimal thermoregulatory response are only achievable on a regular basis during the day in forested environments; whereas in the open, under direct sun exposure, temperature conditions are much hotter warranting limited exposure during peak daytime temperatures to avoid the risk of overheating, even when considering the difference between ambient air temperature and the soil surface temperature. Therefore, it would have been critical for the survival of hominids to escape the extreme temperatures of midday open environments, or at least limit their exposure to the direct sun. From the graphs above, it can be concluded that exposure in the open from about 1000 to 1600 local time would result in a net increase in body temperature due to the heat gained from the environment and through metabolic activity. Conclusions We present the environmental parameter of temperature as a "statistical composite day" whereby the hourly ground surface temperature is estimated over seasonal periods; these are calculated from continuous subsurface measurements of soil temperature using 75 the 1-dimensional heat diffusion equation. We have shown that environmental temperature differences between different microclimatological settings are extremely high: daily maximum soil surface temperatures at low altitude open settings regularly exceed 50°C and periodically exceed 60°C in some open grassland environments. Open settings show large diurnal variation often on the order of ca. 20-30°C. Nearby forested sites have daily maximum surface temperatures in the mid 30°C. At higher elevations, steep spatial differences in ground surface temperatures exist but are much less extreme. In the open, diurnal ranges are on the order of 20°C and daily extremes in the open are regularly in the range of 40-50°C. These observations are relevant for consideration of microenvironments relevant to hominid evolution, because high soil temperatures have been measured for paleosols in the Turkana Basin. 76 77 References Bobe, R and Leakey, MG (2009), Ecology of plio-pleistocene mammals in the Omo- Turkana Basin and the emergence of Homo. The first humans - Origin and Early Evolution o f the Genus Homo, eds Grine FE, Fleagle JG, & Leakey RG (Springer). Bartlett, M. G., D. S. Chapman and R.N. Harris (2006), A decade of ground-air temperature tracking at Emigrant Pass Observatory, Utah, Journal o f Climate 19(15): 3722-3731. Carslaw, H. S. and J. C. Jaeger (1986), Conduction o f Heat in Solids, Oxford: Clarendon Press. Cerling, T. E., N. E. Levin, J. Quade, J.G. Wynn, D.L. Fox, J.D. Kingston, R.G. Klein and F.H. Brown (2010), Comment on the paleoenvironment of Ardipithecus ramidus, Science 328(5982): 1105. Eichna, L. W., C. R. Park, N. Nelson, S. M. Horvath and E. D. Palmes (1950), Thermal regulation during acclimatization in a hot, dry (desert type) environment, The American Journal o f Physiology 163(3): 585-597. Falk, D. (1990), Brain evolution in Homo: The "radiator" theory, Behavioral and Brain Sciences 13(2): 333-381. Feibel, C. S. (2012), A geological history of the Turkana Basin, Evolutionary Anthropology 20(6): 206-216. Feibel, C. S. and J. M. Harris(1991), Palaeoenvironmental context for the late neogene of the Turkana Basin, Koobi Fora Research Project 3: 321-370. Geiger, R., R. H. Aron (2009), The Climate Near the Ground, Rowman & Littlefield, Lanham, MD. Griffiths, J. F. (1972), Climates o f Africa, Elsevier Pub. Co., Amsterdam, NY. Hijmans, R. J., S. E. Cameron and J. L. Parra, (2005), Very high resolution interpolated climate surfaces for global land areas, International Journal o f Climatology 25(15): 1965-1978. Jablonski, N. G. and G. Chaplin (2000), The evolution of human skin coloration, Journal o f Human Evolution 39(1): 57-106. Jablonski, N. G. and G. Chaplin (2010), Human skin pigmentation as an adaptation to UV radiation, Proceedings o f the National Academy o f Sciences o f the United States o f America 107(SUPPL. 2): 8962-8968. 78 Jury, W. A. and R. Horton (2004), Soil Physics, John Wiley & Sons, Hoboken, NJ. Kerslake, D. M. (1972), The Stress o f Hot Environments. Cambridge Univ. Press, Cambridge, MA. McFarlane, W. V. (1968), Adaptation of ruminants to tropics and deserts. Adaptation o f Domestic Animals. E. S. E. Hafez. Philidelphia. Nicholson, S. E. (1996), A review of climate dynamics and climate variability in Eastern Africa. The Limnology, Climatology and Paleoclimatology o f the East African Lakes, Gordon and Breach: 25-56, Amsterdam, The Netherlands. Passey, B. H., N. E. Levin, T.E. Cerling, F.H. Brown and J.M. Eiler (2009), High-temperature environments of human evolution in East Africa based on bond ordering in paleosol carbonates, Proceedings o f the National Academy o f Sciences 107(25): 11245-11249 Potts, R. (1998), Environmentsal hypotheses of hominin evolution, Yearbook o f Physical Anthropology 41: 93-136. Ruxton, G. D. and D. M. Wilkinson (2011), Avoidance of overheating and selection for both hair loss and bipedality in hominins, Proceedings o f the National Academy o f Sciences o f the United States o f America 108(52): 20965-20969. Wallace, J. M. and P. V. Hobbs (2006), Atmospheric Science: An Introductory Survey, Elsevier Academic Press, New York. Wheeler, P. E. (1984) The evolution of bipedality and loss of functional body hair in hominids, Journal o f Human Evolution 13(1): 91-98. Wheeler, P. E. (1991), The thermoregulatory advantages of hominid bipedalism in open equatorial environments: The contribution of increased convective heat loss and cutaneous evaporative cooling, Journal o f Human Evolution 21(2): 107-115. Wheeler, P. E. (1992), The influence of the loss of functional body hair on the water budgets of early hominids, Journal o f Human Evolution 23(5): 379-388. White, F. (1983), The vegetation of Africa, Natural Resources Research 20. White, T. D., S. H. Ambrose, G. Suwa and G. Woldgabriel (2010), Response to comment on the paleoenvironment of Ardipithecus ramidus, Science 328(5982): 1105. White, T. D., B. Asfaw, Y. Beyene, Y. Haile-Selassie, C. O. Lovejoy, G. Suwa and G. WoldGabriel (2009). "Ardipithecus ramidus and the paleobiology of early hominids." Science 326(5949): 64, 75-86. CHAPTER 4 SUMMARY AND CONCLUSIONS In the previous two chapters we presented soil temperature data for the purpose of assessing the spatial variability of soil temperatures in Kenya over a range of elevations and vegetation types. From this work we can arrive at the following conclusions. 1. Variability of soil temperature throughout Kenya is high. The hottest soil temperatures are observed in low-lying grassland and bushland settings, whereas the coolest temperatures are observed in the higher altitude forested settings. 2. In all instances, soil temperatures within forested settings are cooler than those in direct exposure to the sun. 3. The warmest soil temperatures at 25 cm, which is consistent with depths in which soil carbonate formation is observed, are found within the Turkana Basin at Ileret and at the Tana River Primate Reserve. 4. At Ileret the mean annual average soil temperature at 25 cm is 34°C with the hottest temperatures observed being just shy of 39°C during the DJF season 5. At Tana River the mean annual soil temperature at 25 cm is 33°C with the hottest temperatures observed during the DJF season also being just shy of 39°C. 6. A statistical composite day is an effective way of distinguishing the averaged diurnal features such as the daily maximum, minimum, and range. The creation of a composite day for each season clearly illustrates the differences caused by seasonal monsoon cycles driving the long and short rainy seasons. 7. Daily maximum temperature differences between shaded and unshaded environments at low elevations can be as great as 25°C and at higher elevations can be 10-15°C. During the night, when different environmental types come into radiative equilibrium, the temperature difference between them is very small. 8. Temperature variability on daily, seasonal, and annual scales within a forest is much smaller than it is in the open as the forests limit the amount of solar insolation during the day and IR radiation at night. Therefore, diurnal ranges in soil temperature in the forest are ~5°C, whereas they can be as high as 30°C in the open. 9. The temperature difference from the soil surface to the air temperature measured at 2 m above the ground can be quite significant during the day and are quite similar at night. This difference is also dependent upon the type of vegetation cover at the site. At higher elevations, in Nairobi, the difference between soil surface and air temperatures in the open is 10-15°C, whereas it is only 2-5°C in the forest. At lower elevations the difference from the soil surface to 2 m are on the order of 20°C in the open, whereas average maximum temperature of the soil surface in the forest is cooler than the elevation corrected air temperature by ~4- 5°C. 80 APPENDIX A SITE LOCATION AND DESCRIPTION In this appendix we lay out the various site descriptions. The exact location of each site can be found in Table A.1. Meru National Park Meru National Reserve is located on the equator in Central Kenya. The park covers and area of around 900 km2. At Meru, the closest permanent weather station is located at Isiolo, which is 34 mi northwest of Meru National Park. At Isiolo, the mean annual precipitation is 650 mm. The mean annual air temperatures for this location are 23.5°C. Temperatures are highest in February and March and lowest in July and August, however, the deviation is small with a standard deviation of only 1°C through the year. Within this park there were three locations selected including grassland, bushland, and riparian forest zones. In addition, there was one location selected in a bushland community area immediately adjacent to the park. Both of the bush sites at Meru have about 50% canopy cover. The open site at Meru has no notable canopy cover but has dense grass coverage about 1 m high. The forested site at Meru is a dense tropical riparian forest with approximately 96% canopy cover. Kakamega Forest National Reserve Kakamega Forest National Reserve is located in the western highlands of Kenya. Annually the area receives 2000 mm of rain. The closest weather station is located at Busia, which is approximately 85 km east of the park. Mean annual air temperature at the park is 22°C. At this location there were two measurement sites. One site was located in a dense tropical forest and the other was located in open grassland. The open site at Kakamega forest has approximately 3% canopy cover, and acacia trees mainly provide this cover. The forested site at this location is a dense highland forest system; the canopy cover within this forest is 99%. Lake Nakuru National Park Lake Nakuru National Park is located between Mt. Kenya and Kakamega forest. Annually the area receives 1000 mm of rain. Mean annual air temperature at the park is 17.5°. Daily air temperature highs are in the rage of 25°C with the hottest temperatures observed in February and March. At this location two sites were instrumented in 2009 and three additional sites emplaced in 2010. In 2009 we selected one site located in a stand of Acacia xanthophloea with approximately 39% canopy cover and the other from this period was in an open grassland site with no canopy only grass coverage. During May of 2010 we chose to add three more locations, one located in a bushland area, an additional open grassland site located approximately 30 m from the lake margin, and a second forested site within an Olea forest. 82 83 Table A.1 - Latitude, Longitude, and Elevation of Soil Temperature Sites. Site Latitude Longitude Elevation (m) Arabuko Sokoke Brachystegia -3.32153 39.92542 31 Arabuko Sokoke Mixed -3.32155 39.93288 27 Arabuko Sokoke Cynometra -3.32079 39.88726 59 Arabuko Sokoke Disturbed -3.30236 39.9968 25 Ileret Open 4.28808 36.26057 430 Ileret Bush 4.28809 36.26057 440 Kakamega Open 0.3479 34.86915 1572 Kakamega Forest 0.35604 34.860662 1628 Meru Open 0.18012 38.22673 593 Meru Bush -0.07013 38.41288 343 Meru Disturbed -0.28043 38.20503 514 Meru Forest -0.07175 38.42017 330 Mt Kenya High -0.16975 -0.16975 3072 Mt Kenya Low -0.1729 37.15321 240 Nairobi Open -1.35143 36.79628 1691 Nairobi Bush -1.351383 36.796399 1690 Nairobi Forest -1.34836 36.767312 1792 Nakuru Open -0.41755 36.12577 1781 Nakuru Forest -0.417950 36.124970 1785 Shimba Hills Open -4.23394 39.41922 380 Shimba Hills Forest -4.23464 39.41929 390 Tana Forest -1.87652 40.13994 43 Tana Open -1.87615 40.13791 37 Tsavo West Open -2.74728 38.12858 886 Tsavo West Bush -2.74705 38.12828 884 Tsavo East Open -3.36249 38.64526 507 Tsavo East Forest -3.36228 38.64492 507 Tsavo East Bush -3.3625 38.645289 507 Mt. Kenya National Park Two locations were selected on Mt. Kenya. Both are high elevation sites and are located within dense bamboo forests with bamboo canopy coverage of 100%. Air temperatures at these locations are much cooler than the rest of Kenya and differ from each other significantly. Rainfall is also variable because of the high elevations. The closest weather station to Mt. Kenya is Nanyuki, which is unrepresentative of what actual conditions would be on site. Tana River Primate Reserve Tana River Primate Reserve is located in Eastern Kenya along the Tana River. This riparian forest is home to two endangered endemic species of monkey, the Tana River Mangabey and the Tana River Red Colobus Monkey. At Garissa, 170 km north of Tana River Primate Reserve, annual precipitation is 300 mm and the mean annual temperature is 29°C. At this location there were two sites selected, one in an open setting with no canopy or bush cover. The other site was located within the dense tropical riparian forest consisting of approximately 96% canopy coverage. Tsavo National Park Tsavo National Park is located between Nairobi and Mombasa. The park is split into Tsavo East and Tsavo West. At Tsavo West there were two measurement locations: one in open grassland, and the other under the shade of a bush. At Tsavo West the grassland location has no notable tree canopy coverage but has approximately 38% of decent short (0.5 m) shrub coverage and medium (0.5 m) height grass coverage. The bushland site at Tsavo West was located underneath a bush that stands approximately 3 m high that provided about 50% canopy coverage to the site. At Tsavo East there were three sites: open grassland, bushland, and a small isolated forest. The open site had no canopy cover and only short grass; the bush site was located under a dense bush that provides about 40% canopy cover. The forest location was located under a mixture of a large tree and bush coverage that provide a significant amount of cover, approximately 70%. 84 Arabuko Sokoke Forest Reserve Arabuko Sokoke National Reserve is located in Southeastern Kenya. The park lies about 10 km inland from the Indian Ocean and about 20 km southwest of Malindi, Kenya. Arabuko Sokoke Reserve is the largest stretch of costal dry forest remaining in East Africa and is entirely a forested ecosystem. The reserve is comprised of three forest types: mixed forest, brachystegia woodland, and cynometra woodland. All three forest types are dense forests with greater than 70% canopy coverage. MAP in this region is about 1000 mm annually. At the closest meteorology station, Malindi, the mean annual air temperature is 26°C. The warmest daily maximum temperatures, around 30°C, are recorded in December, January and February and the coolest daily maximum temperatures, around 27°C, are recorded in June, July and August. Lows throughout the year are about 22°C. Shimba Hills National Reserve Shimba Hills is located approximately 30 km south of Mombasa. The park is comprised of costal forest ecosystems, open bushlands, and open grassland rages. Next to Arabuko Sokoke, Shimba Hills holds one of the largest costal forests in East Africa. MAP in this region is about 1000 mm annually. At the closest meteorology station (Malindi), the MAAT is 26°C. The warmest daily maximum temperatures, around 30°C, are recorded in December, January, and February and the coolest daily maximum temperatures, around 27°C, are recorded in June, July, and August. Lows throughout the year are about 22°C. At this location, temperature pendent strings were located in a 85 dense costal forest and also in an adjacent open field. The two monitoring locations were at approximately the same elevation of 300 m. 86 APPENDIX B ILERET AND TURKWEL WEATHER STATIONS: THE FIRST YEAR OF DATA In May of 2010 two permanent weather stations were installed at the Ileret and Turkwel field camp locations. These weather stations were installed in conjunction with the Turkana Basin Institute (TBI) because of the lack of climate monitoring coverage in the area. Previously, the only station that was recording any climate data within and surrounding the Turkana Basin was located at the Lodwar airport and was operated by the Kenya Meteorological Service (KMS). In addition to the limited data coverage, the ability to obtain the data from KMS is somewhat of a challenge and quite expensive. The weather stations installed at both locations are identical weather stations made by Onset Computer Corporation. A list of components and part numbers can be found in Table B.1. As stated before, the weather stations are located at the Turkwel and Ileret field stations and are situated on either side of Lake Turkana, and separated by approximately 135 km. The coordinates for the stations and elevations are given in Table B.2. The TBI staff maintains the weather stations and the data are downloaded on a weekly basis (typically Sunday) and the files are transferred via email in a proprietary format (.HOBO). Sensor Descriptions Each weather station was comprised of 12 sensors. Below is a brief description of the sensors and their locations. An annotated picture of the weather station is presented in Figure B.1. • Soil Temperature - There were three soil temperature thermistors buried adjacent to the weather station. The thermistors were buried at 0.05 m, 0.15 m, and 0.25 m to correspond with the soil temperature monitoring project. The soil in both locations was mostly sandy with very little vegetation. • Soil Moisture - There were two soil moisture sensors buried at 0.05 m and 0.25 m. These sensors were buried adjacent to the soil temperature sensors. • Temperature/RH - The temperature and relative humidity probe was located approximately 2 m off the ground and attached to the main mast of the tripod. It was shielded using an unaerated radiation shield. • Barometric Pressure - The barometric pressure sensor was located directly outside the data logger housing and was attached directly to the main tripod mast. • Wind Speed - Wind speed was measured using a cup anemometer located atop the main mast of the tripod. • Wind Direction - Wind direction was measured using a wind vane located adjacent to the cup anemometer. The wind vane was oriented true north. • Solar Radiation - Incident solar radiation was measured using a silicon pyrometer located atop the main tripod mast. Because of the geolocation of the weather stations this sensor was oriented so that no shadowing will occur from either the tripod mast of the wind sensors. 88 • Rain - Rain amount was measured using a tipping bucket rain gauge that is located on the main mast of the tripod opposite the temperature sensor. Data Processing Data Processing for the weather stations is done using two programs, the first is a proprietary software, HOBOware® Pro and the second is the Interactive Data Language (IDL) programming language. The proprietary HOBOware® software is required in order to interface with the weather station and download data. Additionally, basic plots and statistics can be obtained with this software; however, it is not very useful beyond basic functions. From HOBOware® the data must be exported as a comma separated file in order to import the data to excel or conduct further processing and quality control. Further processing and quality control is conducted using IDL. These programs output eight data sheets of varying timescales. Fifteen min, daily, monthly, and annual averages are updated with new data as it becomes available. In addition to the data tables, the programs output weekly "quick look" images for each sensor on the station; this gives the user the ability to observe any misbehaviors or faulty sensors. First Year of Data and Results Plotted in Figures B.2 and B.3 are the daily air temperature maximum, minimum, average, and range for Ileret and Turkwel, respectively. It is quite incredible that at both locations the average daily temperature, calculated as the average of the daily high and daily low, is frequently in excess of 30°C. Annually, the average daily temperatures for Ileret and Turkwel compare quite well at 31°C. Although daytime temperatures are 89 90 Table B.1 - Parts list for the weather stations at Ileret and Turkwel Field Stations. Part Number Description Price (US$) SOLAR-6W 6 Watt Solar Panel 149 BHW-PRO-CD HOBOware Pro Mac/Win 99 U30-NRC-VIA- 10-S045-000 USB Weather Station Data Logger 527 S-BPB-CM50 Barometric Pressure Smart Sensor 249 S-LIB-M003 Solar Radiation Sensor (Silicon Pyranometer) 210 S-RGA-M002 .01" Rain Gauge Smart Sensor (2m cable) 410 S-THB-M008 12-bit Temperature/RH Smart Sensor (8m cable) 195 S-SMC-M005 S-SMC-M005 - Soil Moisture Smart Sensor 139 S-TMB-M006 12-Bit Temp Smart Sensor (6m cable) 105 S-WSET-A Wind Smart Sensor Set - S-WSET-A 560 RS3 Solar Radiation Shield 59 M-TPA-KIT HOBO Weather Station 3-Meter Tripod Kit 275 91 Table B.2 - Latitude, longitude, and elevation of the weather stations. In November of 2011 the weather station at Turkwel was moved because of the construction of a new building. The data contained within this discussion is all from the old location.________ Site Latitude Longitude Elevation (m) Ileret 4.28767 36.26105 438 Turkwel (Old Location) 3.13957 35.86455 460 Turkwel (New Location) 3.14169 35.86354 467 92 Wind Speed and Direction Sensor Figure B.1 - Annotated image of the weather station that is located at Turkwel Field Station. The Station located at Ileret is identical. 93 -High -Low Range -Average Figure B.2 - Daily air temperature high, low, range, and average at Ileret on the east side of Lake Turkana from June 2010 through May 2011. Temperature °C 94 -High Low -Range - Average Figure B.3 - Daily air temperature high, low, range, and average observed at the Turkwel weather station on the west side of Lake Turkana from June 2010 through May 2011. consistently quite warm, it is the nighttime lows that keep the daily average so high as they rarely fall below 25°C. It is easily seen from these figures why the Turkana Basin is amongst the hottest 1% of places on earth [Hijmans, 2000]. In addition to being one of the hottest places on earth, the Turkana Basin, in contrast to many other tropical environments, is very dry [Griffiths, 1972]. During the first year of these stations being operational there was a mere 130 mm of rain that fell at Ileret. It was much drier at Turkwel with a pitiful 22 mm of rain that fell over that year. Contrast this with other areas within Kenya that receive over 2000 mm of rain each year. The dryness of the basin can also be seen in the soil moisture and temperature measurements. The daily averaged soil moisture measurements are presented in Figure B.4 and B.5, for Ileret and Turkwel, res