Simsphere is a time-dependent, one-dimensional, two-stream model that accounts for the interaction between the soil, the vegetation, and the atmosphere over a 24-h day. One of a class of so-called two-stream models, Simsphere simulates fluxes over a surface area consisting of fractional parts of vegetation and bare soil and blends the fluxes in the atmosphere weighted by the fractional vegetation cover. As such it simulates the fluxes of heat and moisture exchanges between these three layers in parallel over bare soil and vegetated surfaces, while allowing changes to occur in the atmospheric temperature, dewpoint, and wind velocity during daylight and at night and changes in temperature and water content in the substrate. Simsphere also allows for the flux of carbon dioxide from the atmosphere to the plant canopy. Simsphere’s basic structure is encapsulated in Fig. 1. Figure 2 is an example of a graph made from its output, showing variations in the various flux components during a 24-h simulation.
Basic structure of our model.
Citation: Bulletin of the American Meteorological Society 102, 12; 10.1175/BAMS-D-20-0296.1
A graph of an example of output from the bare soil version of the Simsphere model. Net radiative flux (NETRAD) is plotted in blue, along with the latent heat flux (LE; green), the sensible heat flux (SENS; red), and the substrate heat flux (GRF; gold).
Citation: Bulletin of the American Meteorological Society 102, 12; 10.1175/BAMS-D-20-0296.1
Many so-called soil–vegetation–atmosphere–transfer (SVAT) models have been created and published in the literature, but it is not our intention to present a literature review other than to cite a few relevant papers specific to the operation of Simsphere. The vast majority of these references (Cosby et al. 1984; Anagnostopoulos et al. 2017) pertain to the construction and validation of Simsphere and are included for the sole purpose of aiding the user in understanding and operating Simsphere. Simsphere has been used extensively as a research tool or in formal coursework at seven different universities and one institute in three countries; this history is summarized in Table 1.
Simsphere in academia: universities, administering faculty, and abbreviated course names for undergraduate and graduate curricula.
What distinguishes Simsphere, besides being the product of 30 years of continuous development and validation, is that it is now being made freely available for download on the internet, both as an executable file and as a code that can be compiled and modified by users. Accompanying the model on its website are many documents describing the model structure, its access to the user, a flow diagram of its structure, input, and output, as well as relevant journal references or publications (including this paper) describing parts of Simsphere and its various applications.
The purpose of this paper is to inform the scientific community of the availability of Simsphere and to briefly describe its structure in order that the user can access and download the model from the internet. First, we briefly describe some of the model structure and capabilities. Next, we describe the website and show how the model can be downloaded and run within the local user environment. Finally, we describe the workbook.
Structure of Simsphere
Consisting of approximately 50 pages of Fortran-90 code, Simsphere is not able to be fully described here or in any single document or publication. Instead, we will list only a few of the most relevant and useful papers in order to facilitate an understanding of the model and its operation on the web. Many of the mathematical details of the model, including some validation, can be found on the Simsphere website and in the following papers on this site (to be discussed later): Gillies and Carlson (1995), Lynn and Carlson (1990), Carlson and Lynn (1991), Olioso et al. (1996), Grantz et al. (1999), Carlson (2007), and Petropoulos et al. (2009, 2013, 2015). The Simsphere website (https://simsphere.ems.psu.edu/) also contains, besides the aforementioned references to published papers, many useful documents: tables of input and output variables, an extended description of the model, the uses of the input variables, and a sample output table. These are further referred to below. Perhaps the most extensive document is a workbook, which can serve as a tutorial for the user and as a teaching tool. The workbook is also described below.
Unlike three-dimensional atmospheric models that require large and carefully filtered datasets for initialization, the one-dimensional Simsphere is far less complex to execute. After downloading and building (if necessary) the model, the latter can be run immediately using the default initial conditions built into the model. These settings, which pertain to a summer day in the Midwest of the United States, are found in the i_model.json input data file. This file defines some basic model parameters such as the latitude, longitude, and elevation at which the model performs its integrations over time (xlat, xlong, station_height), and the date and time at which the model begins integration (year, month, day, strtim) and ends. An initial atmospheric sounding is defined at 12 pressure levels for temperature and dewpoint depression, and at 11 altitudes for wind direction and speed. Additionally, there are parameters that define the properties of the ground, vegetation, and air. For example, the initial amount of water content in the soil is defined by “f” (surface layer moisture availability) and “fsub” (root zone moisture availability). The fraction of the ground covered by vegetation is defined by “frveg,” and the areal coverage of leaf surfaces by “xlai” (the leaf area index). Many other parameters can be varied as well, such as the altitude, azimuth, and elevation angles of the surface; these are defined in two tables. The next three sections briefly describe the main components of Simsphere.
Substrate layer.
Water content is specified as percentage of field capacity. It is held constant in the “root zone,” a soil layer of indeterminate depth, but water can flow to or from this layer from or to a surface layer whose water content can vary with time according to the evaporation rate and the rate of flow to or from the root zone. Water can also flow from the surface soil layer to the atmosphere through the bare soil surface or through the plants from the root zone, but transpiration is regulated by a series of soil, plant, and leaf resistances (e.g., the stomatal resistance). Heat flow occurs into or out of the soil from the atmosphere, allowing the substrate temperature to change between the surface and a lower level whose temperature is fixed at the monthly average. Evaporation can occur with or without plants from the soil surface to the atmosphere. Different types of soils can be chosen by the user from a soil table, each soil having its own set of properties (thermal diffusivity and conductivity), as defined in Cosby et al. (1984).
Plant layer.
Five different plant types can be chosen by the user from a table of plant values labeled by the species. For each plant a set of parameters, such as the minimum stomatal resistance and the critical leaf water potential, is specified. Although the parameters are roughly known for these five varieties of plants, the table can be modified or expanded by the user, if desired, in order to accommodate different parameters or different plants. Water flows from the root zone through a series of resistors (soil,root, stem, leaf) to the leaves and thence into the atmosphere or to the surrounding interplant spaces as transpiration. Heat and moisture flow from the leaves to the air above the plant canopy and into the interplant spaces. In addition to transpiration fluxes, carbon dioxide fluxes over the plant canopy are calculated. Ozone fluxes to the ground and plant canopy are calculated, given a fixed atmospheric O3 concentration.
A somewhat esoteric concept of the plant hydrology (generally unfamiliar to meteorologists) is the use of water potential (for soil, stem, and leaf). This parameter allows the stomatal resistance in the leaf to be mathematically tied continuously to the root zone, stem, and leaf water flow. Three stomatal resistance models are available for the user’s choice. However, we highly recommend that the Lynn and Carlson (1990) version be used in view of its extensive application. Simsphere also accounts for plant water storage in a capacitance module, but we do not recommend it to be called unless the user has a thorough understanding of its mathematical structure and physics, as described by Carlson and Lynn (1991).
The plant table specifies certain constants that govern transpiration—minimum stomatal resistance, critical epidermal (leaf) water potential, stem resistance, etc. Values for some of these parameters were based on guesswork, although they are somewhat more reliable for corn, soybean, and cotton for which simulations with Simsphere were made in conjunction with field measurements.
Heat and moisture fluxes to the atmosphere are calculated for the bare soil fraction and for the vegetated fraction. The two flux streams for vegetation and bare soil are melded above the canopy according to the amount of fractional vegetation cover, which is specified by the user. Setting the fractional cover to zero activates only the bare soil component. The vegetation component is inactive at night (solar sunset). An example showing transpiration fluxes over a soybean canopy as calculated by Simsphere and another model is shown in Fig. 3 with a comparison to surface measurements (Olioso et al. 1996).
Diurnal evolution of photosynthesis and transpiration simulations over soybeans on selected days near Avignon, France. Simsphere results are shown by the dashed line and those from a model by Olioso et al. (1996) by the continuous line. Experimental data are represented by asterisks.
Citation: Bulletin of the American Meteorological Society 102, 12; 10.1175/BAMS-D-20-0296.1
Atmospheric layer.
Incident solar and longwave radiation fluxes at the surface are also exchanged between plant, atmosphere, and soil. Albedo for both plant and soil can be either calculated internally or specified by the user. Fluxes of heat and water vapor from the vegetation canopy and from the soil beneath the plants or surrounding the plants (from the bare soil fraction) are passed into the atmosphere, causing changes in its temperature, dewpoint, and wind velocity. These changes occur initially within a surface layer forming a mixing layer that grows throughout the day. As such, a realistic evolution of temperature, specific humidity, and wind velocity evolve to a height just above the top of the mixing layer.
Nocturnal cooling occurring after radiation sunset results in a shallow cooling layer topped with an inversion that grows with time. Intermittent turbulent episodes can occur at night according to a Richardson number criterion. In the absence of such turbulence, Simsphere is capable of forming a low-level jet during the night. Latent heat fluxes continue weakly after sunset, usually disappearing shortly thereafter. Sensible heat fluxes are generally slightly negative at night except in transient turbulent episodes wherein small increases in surface wind speed and sensible and latent heat fluxes can occur.
Downloading and operating the model from the website
Simsphere exists on the website https://simsphere.ems.psu.edu/ as a source code that can be downloaded and either compiled from a source code or operated as an executable binary code that can be run without compiling. Clicking on the “Model” link will display a brief description of the model along with additional links pointing to a more detailed Simsphere overview with a “Quick Start Guide,” and a community mailing list for Simsphere-related topics of discussion.
The native build environment for the model is Linux, but it can be built on Windows with the aid of a package such as MSYS2, which installs a Linux-style environment on Windows. The simplest way to download and run the model on one’s computer is to follow the instructions in the Quick Start Guide and use the default input files initially. Those instructions will explain system requirements, the download procedure, environment variable settings, and the build and run commands. In addition to the model, a variety of auxiliary files relevant to understanding and executing Simsphere are found under the
“Resources” link. This includes the “Technical Manual” (detailed documentation of the model theory of operation), input model parameter lists and parameter descriptions, listings of the codes, a sample output, a bibliography, and the Simsphere workbook. The bibliography contains, in addition to this paper, an abridged collection of relevant published papers in which Simsphere has been evaluated, validated, or applied to specific problems.
All control files and initialization data necessary to run Simsphere are included with the distribution package and are set to default values so that the model can be run immediately after installation. The primary control file is called “i_model.json.” This file is read from the work area created during installation and contains numerous parameters that allow the user to control what components of the model are used during execution, but also establishes the initial conditions of those components. These parameters include fundamental definitions such as the location, time of year, average terrain slope, and the initial sounding values of temperature, humidity, and wind for which the 1D model will begin execution and the start and end times. In addition, there are parameters that define soil moisture conditions, the presence and nature of surface vegetation, cloud cover, albedo, etc. Two other files, “soils.dat” and “veglut.dat,” allow detailed control of soil and vegetation properties, respectively. The soils file predefines different kinds of soils over which the model can run, and if desired, the user can add additional soil types. Similarly, the vegetation file contains several predefined vegetation types that can be extended as well. These files are located in the “data” directory of the Simsphere distribution tree.
After running the model, two output files called “o_model.dat” and “o_model.json” are produced. The “o_model.dat” file is a text-oriented output file containing a listing of the input parameters for reference, followed by the model results at half-hour intervals. The results are printed in a numerical fixed text format and contains most of the fundamental model parameters illustrating SVAT interactions over time. Additional output can be added programmatically, if desired. Although Simsphere does not produce output graphics, the JavaScript Object Notation (JSON) output (which contains only model results) is useful for loading into graphing software to produce visually interpretable results. The JSON format is also useful for those wishing to integrate Simsphere with other tools, such as Jupyter notebook or a JavaScript web interface.
A simple Simsphere experiment
The amount of moisture available in the surface layer of nonvegetated soil has a large impact on the effect of daytime solar heating on the air temperature above that surface. When little moisture is available, most of the absorbed solar energy is used to heat the air, but when the soil surface is moist, some solar energy is used to evaporate liquid moisture into the air as water vapor releasing latent heat. But how much does this affect the air temperature? Using Simsphere, we can estimate this effect.
The effect of removing most of the surface available moisture is clearly seen in the reduced latent heat flux (almost half), the increased sensible heat flux (more than double), and the increase in the 10-m air temperature by 3.5°C. These kinds of effects would be observable, for example, when comparing temperate to dry climates or when a normally temperate area experiences drought conditions.
The workbook
The workbook is a step-by-step journey through the fundamental components of the Simsphere model starting with the simplest bare soil case and adding increasing SVAT complexities as additional model components are brought to life. To those unfamiliar with SVAT models, the workbook might be the best means for understanding it.
Consisting of 13 chapters preceded by a lengthy introduction, the workbook serves as both a primer for Simsphere and as a teaching tool. Early chapters address the effects of solar and longwave radiation, wind speed, humidity, and temperature on the surface fluxes, as represented in Fig. 2, and progresses with increasing complexity such as wind and boundary layer formulations. Later chapters include vegetation and the effects of water stress on the behavior of plants. An example of the use of Simsphere illustrating water stress over a soybean canopy is shown in Fig. 3. The ability to simulate the decline in transpiration with time and the evolution of the so-called midday stomatal closure is highlighted in this figure.
The last few chapters include some rather complex but schematic plant physiology, showing how atmospheric and soil factors influence plant behavior, including carbon dioxide fluxes. Although the plant formulations are highly simplified (as compared with their actual physiology), they are nevertheless complex and able to simulate subtle aspects of plant behavior, such as the “feed forward” effect in which transpiration decreases with increasing dryness of the atmosphere. The best way to understand Simsphere is to run a simulation with it, and the best way to do that is to read and follow the workbook as it progresses through its increasing complexity.
The purpose of the workbook is to act both as a tutorial for the user and as a potential teaching tool. Each chapter concludes with a series of proposed simulations designed to elucidate various points raised in the body of the chapter. Two levels of simulations are presented. Level 1 proposes some simple simulations while simulations proposed in level 2 require the user to be creative in achieving a satisfactory result, which may not be attained without some effort and, in some cases, that effort may yield ambivalent results. Mathematical formulations in the workbook are limited to an essential schematic level, although those same formulations in the code are much more elaborate. Those familiar with the mathematics of such models will recognize the simpler versions. Those unfamiliar with the mathematics will still be able to appreciate the results obtained with more detailed formulations in the model and may be able to recognize them in the code.
Concluding remarks
Simsphere is offered free to anyone who wants to use it and for any application that one can devise, whether for meteorological, engineering, agriculture, or societal purposes. Users are free to operate Simsphere as it stands, modify it, or extract portions for their own needs. For those wishing to understand the model, we highly recommend studying the various documents available on the website (such as the workbook), as well as reading some of the published papers cited therein. A structure developed over 30 years and its complex physics contained within cannot be understood overnight. One purpose of making this model available to the public is to forestall future efforts to reinvent the wheel, so to speak. As such, anyone wishing to develop an SVAT model will recognize its precedent and will be able to employ its various components for other purposes. Although the model will likely receive occasional limited and controlled updates on the Simsphere website, one should feel free to download the Simsphere code and to restructure the code as seen fit, even using its components and formulations in other models.
Future plans for the model include a graphical interface to adjust model parameters and visualize output. Containerization of the application is also planned. Contributions from the community toward these plans and other general improvements to the model software are welcome. We ask that if anyone makes substantial changes in the Simsphere model itself, or adapts Simsphere for some other purpose, they please inform us of advancements that have been made, if only for the sake of courtesy and to share with us Simsphere’s legacy.
A word of encouragement.
Thirty years of continual use, testing, and validation attest to the robust nature of Simsphere. Model crashes, should they occur, almost always signal that the user has entered untenable conditions for Simsphere to execute, for example, specifying a vanishing root zone soil water content or an unrealistic surface roughness length. We have found that seemingly bizarre results often signify either that the user has set unreasonable initial conditions or that an unexpected but real phenomenon is being revealed.
Much can be added to the model by the creative user. Such additions might consist of various types of graphical output, for example, an animated temperature–dewpoint skew T sounding, or the inclusion of different types of severe storm indices that would change as the sounding evolves throughout the day. We wish the user good luck and success with Simsphere.
Acknowledgments.
Over the years many students and colleagues have contributed to Simsphere development. Besides the current authors of this paper, we need to thank Rob Gillies, Barry Lynn, David Ripley, David Grant, Albert Olioso, Kell Wilson, Dev Niyogi, and Odile Taconet for their assistance. We are also grateful to our College of Earth and Mineral Sciences under Dean Lee Kump and the Department of Meteorology and Atmospheric Science under its head, David Stensrud, for allowing coauthors, Art Person and Tom Canich, to devote some of their precious time to this Simsphere project. Graphics and Design of State College, Pennsylvania, built the website.
References
Anagnostopoulos, V. , G. P. Petropoulos , G. Ireland , and T. N. Carlson , 2017: A modernized version of a 1D soil vegetation atmosphere transfer model for use in land surface interactions studies. Environ. Model. Software, 90, 147– 156, https://doi.org/10.1016/j.envsoft.2017.01.004.
Carlson, T. N. , 2007: An overview of the “triangle method” for estimating surface evapotranspiration and soil moisture from satellite imagery. Sensors, 7, 1612– 1629, https://doi.org/10.3390/s7081612.
Carlson, T. N., and B. H. Lynn , 1991: The effects of plant water storage on transpiration and radiometric surface temperature. Agric. For. Meteor., 57, 171– 186, https://doi.org/10.1016/0168-1923(91)90085-5.
Cosby, B. J. , G. M. Hornberger , R. B. Clapp , and T. R. Ginn , 1984: A statistical exploration of the relationships of soil moisture characteristics to the physical properties of soils. Water Resour. Res., 20, 682– 690, https://doi.org/10.1029/WR020i006p00682.
Gillies, R. R., and T. N. Carlson , 1995: Thermal remote sensing of surface soil water content with partial vegetation cover for incorporation into climate models. J. Appl. Meteor., 34, 745– 756, https://doi.org/10.1175/1520-0450(1995)034<0745:TRSOSS>2.0.CO;2.
Grantz, D. A. , X. Zhang , and T. N. Carlson , 1999: Observations and model simulations link stomatal inhibition to impaired hydraulic conductance following ozone exposure in cotton. Plant Cell Environ., 22, 1201– 1210, https://doi.org/10.1046/j.1365-3040.1999.00486.x.
Lynn, B. H., and T. N. Carlson , 1990: A stomatal resistance model illustrating plant vs. external control of transpiration. Agric. For. Meteor., 52, 5– 43, https://doi.org/10.1016/0168-1923(90)90099-R.
Olioso, A. , T. N. Carlson , and N. Brisson , 1996: Simulation of diurnal transpiration and photosynthesis of a water stressed soybean crop. Agric. For. Meteor., 81, 41– 59, https://doi.org/10.1016/0168-1923(95)02297-X.
Petropoulos, G. P. , T. N. Carlson , M. J. Wooster , and S. Islam , 2009: A review of Ts/VI remote sensing based methods for the retrieval of land surface energy fluxes and soil surface moisture. Adv. Phys. Geogr., 33, 224– 250, https://doi.org/10.1177/0309133309338997.
Petropoulos, G. P. , T. N. Carlson , and H. Griffiths , 2013: Turbulent fluxes of heat and moisture at the Earth’s land surface: Importance, controlling parameters, and conventional measurement techniques. Remote Sensing of Energy Fluxes and Soil Moisture Content, G. P. Petropoulos , Ed., Taylor and Francis, 3– 28.
Petropoulos, G. P. , M. R. North , G. Ireland , P. K. Srivastrava , and D. V. Randall , 2015: Quantifying the prediction accuracy of a 1-D SVAT model at a range of ecosystems in the USA and Australia: Evidenced towards its use as a tool to study Earth’s systems interactions. Geosci. Model Dev., 8, 3257– 3284, https://doi.org/10.5194/gmd-8-3257-2015.
