Research by Niklaus E Zimmermann in the SERDP-project:

Emerging and Contemporary Technologies in Remote Sensing for Ecosystem Assessment and Change Detection on Military Reservations

Principal Investigators: P.T Tueller, R.D. Ramsey, T.D. Frank, R.A. Washington-Allen, S. Tweddale, R.S. Karalus, C.T. Hunsaker, L.E. McCarthy, T.G. Van Niel)

Overall goal of the study

The objectives of this project are to: stratify the landscape of individual military ranges using contemporary and emerging remote sensing technologies; identify the fundamental vegetation and soil attributes of military ranges as they relate to plant succession; establish ecosystem response and recovery in relation to disturbance (land use) through retrospective studies with spatially-explicit spectral-based indices; identify the spatial, spectral and temporal attributes of remote sensing systems necessary to identufy ecotones, and to distinguish along environmental and disturbance gradients, and lastly; develop methods for scaling indices between coarse and fine resolution imagery.

 
 
 

Objectives of the part of Niklaus E. Zimmermann

The objectives of my part of the SERDP-projects are:
  • To conceptually develop high-resolution models (temporally and spatially) for site water balance and for soil wetness.
  • To apply the model to SERDP study areas.
  • To assess the power of this model to improve the SAVI index.
  • To simulate a maximum projected leaf area index (PLAI) for the ecosystems concerned, based on the Hydrological Theory.

    •  

    Introduction

    The research described here is based on the biophysical modeling procedure developed for the Shoshone National Forest. For both projects, the assessment of some sort of soil moisture estimate was/is important. For the SERDP project, however, the procedures and models had to be refined and developed further. Here, a more detailed assessment is necessary in order to distinguish between the soil wetness ans site water balances of individual years, months, and even days. Thus, we need to switch from using lomg-term averages for climatic variables to daily values in order to enable such calculations. Some of the programs and routines developed in the Shoshone project can be used in the present model effort as well, some others need to be adapted or newly developed. Below is a description of concept to achieve the goal of modeling site water balance and soil wetness at high spatial and temporal resolution. All programs developed in this project are made available here. You can download and use these programs. You may alter, integrate them, or adjust them to you needs. But be aware that this is an ongoing project, and that we do not guarantee that they are error-free. We are still adjusting some of the codes, and not all of the manuals have been finalized.
     
     

    1. A concept for modeling site water balance and soil wetness

    Site water balance (SWB) is modeled as a function of climatic and pedological drivers. Soil bucket size and permeability, potential evapotranspiration, and precipitation are combined to calculate the spatially-explicit site water balance similar to the approach first employed by Grier and Running (1977). Beginning with the first month in fall when precipitation exceeds potential evapotranspiration (after a possible drought period), the difference between precipitation and potential evapotranspiration is summed for 12 months. The running sum is never allowed to exceed the bucket size, and water in excess of the bucket size is presumed to run off. Additional water loss (to the groundwater) is simulated as a function of soil permeability. When potential evapotranspiration begins to exceed precipitation the difference is subtracted from water in the bucket, often achieving significant negative values over the course of a year. Site water balance is an estimate of the water available to plants during the course of a year. It integrates both climatic and soil parameters. The method differs from Grier and Running (1977) in the determination of the appropriate water year, in not assuming that soils begin the water year with full recharge, and in adding soil permeability to the simulated processes.

    Soil wetness (SW) is calculated as a by-product of the site water balance model. The principal deviation from the SWB model is that the atmospheric water balance (P-Etp) is not calculated after the soil is either fully recharged or empty. Thus, no negative values are totaled, and the SW reflects the actual wetness of the soil.

    At this stage of the project, we do not integrate lateral flow of water into our models. By this, the present model differ significantly from the TOPMODEL approach. However, we believe that incorporating the spatial pattern of precipitation, global solar radiation, and temperature are increasingly important, if soil wetness estimates have to be calculated over large areas.
     
     

    2. Application of the proposed models

    The simulated distribution of soil wetness is primarily suitable for SAVI calculations where information of the background soil signal have to be weighted by actual soil wetness at the time the satellite images are/were recorded. However, the SW model does not reveal any information on how dry or wet a specific pixel is over the integral of a whole year. Therefore, the SWB model is better suited to assess vegetation structure, biomass, and PLAI. Indeed, a very high correlation has been discovered between SWB and PLAI (e.g. Grier and Running, 1977; Gholz, 1982). Research on this subject led to the development of the hydraulic equilibrium theory (Nemani and Running, 1989). This theory now is the basis of a suite of ecosystem models that simulate leaf area dynamics, photosynthesis, and carbon allocation (e.g. Running and Gower, 1991; Runing, 1994). One of the primary constraints built in such models is that the total ecosystem leaf area never is allowed to exceed the maximum PLAI as defined by the theoretical SWB-PLAI relationship. This maximum possible PLAI can be viewed as the ecosystem potential under undisturbed conditions.

    A disturbed ecosystem will deviate more or less significantly from this potential according to the severity of the disturbance. Since SAVI and other NDVI-related indices give a close approximation of actual ecosystem PLAI, we will be able to calibrate the satellite signal as PLAI for undisturbed sites. Here, we expect a high correlation with the simulated site water balance. Alternatively, we hypothesize to find significant deviations in remotely sensed PLAI from SWB-derived potential leaf area indices for disturbed sites.
     
     
     

    3. Methods

    Below is a brief summary of details of the methods we use to generate the required biophysical modeling environment, specifically,  to generate the site water balance and soil wetness maps. Most of the procedures are calculated in ArcInfo. Some require the application of specific programs, mostly written in Fortran. You can check out the programs and the manuals, and/or you can download them for your own use. Again, be aware that this is an ongoing project, and that we do not guarantee that they are error-free. We are still adjusting some of the codes, and not all of the manuals have been finalized.
     

    3.1. Temperature

    We have agreed with the Utah Climate Center (UCC; Dan Danserau, pers. comm.) to get maps for daily minimum and maximum temperature for the SERDP study sites (1971-present). The maps are generated based on a 500m DEM and data from meteorological recording stations. To enable spatial interpolation, the recorded values were projected to sea level using standard lapse rates. A minimum curvature approach was used to spatially interpolate the station’s values. Finally, the interpolated sea-level temperatures were re-projected to actual elevation using the same lapse rate and a 500m DEM. This spatial resolution is too coarse for our purpose since we operate on a 30m scale. We, thus, will project the temperature maps back to sea level and re-project again to actual elevation using the same lapse rates but a 30m DEM. An alternative source of daily minimum and maximum temperature is the ongoing efforts based on the DAYMET program (Thornton et al., 1997). Large scale maps based on this innovative model will be available by August 1999 (Peter Thornton, pers. comm.).
     

    3.2. Precipitation

    The agreement with UCC includes daily precipitation maps (1971 – present). Since precipitation does not vary at as fine of a spatial resolution as temperature, there is no need to re-project precipitation maps to a 30m resolution. Previous studies suggest that coarser resolutions would be even more appropriate for precipitation (Daly et al., 1994; Thornton et al., 1997).
     

    3.3. Solar radiation

    Considerable efforts have recently been made to simulate solar radiation accurately on large spatial scales in a GIS environment (Dubayah and Rich, 1995; Rich et al., 1995; Kumar et al., 1997). However, these methods only are available in a general form for clear-sky radiation. In order to simulate soil wetness as accurately as possible, we need to incorporate the atmospheric transmittance as well as elevational lapse rates of transmittance. Both can be achieved through the analysis of the hourly solar radiation data as recorded in the SAMSON network (see also NSRDB) and provided by the National Climate Data Center (NCDC). Thus, we will update these solar radiation programs to include sea level actual transmittance and elevational lapse rates. Transmittance is analyzed in a FORTRAN program (see sum_tran.for). We will calculate direct and indirect solar radiation in bi-weekly intervals to save computing time. However, we will switch to a daily interval for the last two weeks before the satellite images were recorded in order to increase the temporal accuracy at the recording date..
     

    3.4. Potential evapotranspiration

    Potential evapotranspiration is calculated using the empirical equation of Jensen and Haise (1963) which was derived from data of the arid western United States. This method is based on daily values for solar radiation and temperature:
     
     ETp = [Rs / 2450 * ((0.025 * Ta) + 0.08)]


    where: ETp = mean daily potential evapotranspiration (mm/day), Rs = daily total solar radiation (kJ/m2/day), Ta = mean daily air temperature (°C). Comparative calculations based on the Solar Thermal Unit method (Caprio, 1974) and the Turc method (Turc, 1963), both based on the same independent parameters, revealed very similar results. The empirical formula of Jensen & Haise is selected because it was especially designed for the arid Western States, and revealed the most likely values. Again, we will stick with the bi-weekly time interval.
     

    3.5. Soil properties

    In order to calculate the soil bucket per pixel, we need to have information on the soil depth (m) and the water holding capacity (mm/m). We derive this, as well as soil permeability, from USDA Soil Surveys and the respective descriptions. For locations where a soil survey is not available, we will employ a topography and geology based approach to assess the respective soil properties. This method is described in Roberts et. al (1993) and under http://www.wsl.ch/staff/niklaus.zimmermann/biophys.html.
     
     
     

    4. Progress report:  March 31st, 1999

    I have performed a series of tasks during the last quarter: namely I (1) checked the availability of successfully operating (and published) climate mapping programs, (2) generated and/or acquired DEM’s for the study sites, (3) tested, corrected and improved available programs to calculate direct and diffuse solar radiation, (4) developed and tested a program to read and summarize (daily, bi-weekly, monthly and yearly) the hourly SAMSON database which contains a series of measured solar radiation data (as well as other climate variables), (5) tested and adjusted ArcInfo routines to calculate potential evapotranspiration (Etp), site water balance (SWB), and soil wetness (SW), (6) developed a spatial database to link additional soils information to the soil maps, (7) written up the manuals and info-files necessary to learn about using these routines and programs, and (8) created an web-based environment to make all information publicly available and downloadable.
     

    4.1. DEM’s for the study areas:

    All DEM’s have been mosaiced and are available in a 30m spatial resolution. The Camp Williams DEM is derived from USGS 7.5’ quads. A series of 28 quads were patched together. Narrow gaps of missing data between adjacent quads were filled by moving averaging techniques. Three missing quads were filled in from the 90m DEM (USGS – 1° quads). Unfortunately, there is a rather big difference in the average elevation per quad between the 30m and the 90m DEM. The Fort Bliss 30m DEM was organized by Robert Washington-Allen.
     

    4.2. Testings, corrections, and improvements of a solar radiation program:

    There are two ready-to-use ArcInfo AMLs available (SHORTWAVE, SOLARFLUX) that allow calculating solar radiation on large grids. Both programs allow the user to include overshadowing (by high adjacent peaks) and to adjust the atmospheric clear-sky transmittance due to various sun angles. One of the two (SOLARFLUX) can cope with the attenuation of radiation due to cloudiness or overcast skies (by simply reducing the transmittance manually), but neither of the two actually includes an elevational lapse rate for transmittance. Some of these features have been included in the DAYMET routine, the code of which is currently not available.

    The program solarflux.aml (Rich et al., 1995) runs smoothly but rather slow, and generates plausible outputs, while shortwave.aml (Kumar et al., 1997) is more than 3 times faster, but calculated erroneous values in winter months for steep south-facing slopes (with „correct" values in all other situations). I found and corrected the bug (see shortwavc.aml). Further tests revealed that besides this, the two programs perform basically the same calculations, the difference is only in the efficiency and in the user interface. I chose to base further development on SHORTWAVE, since it is computationally more parsimonious. I added several features that include: (1) to add an elevational lapse rate for clear-sky transmittance, (2) to add a subroutine that enables reading bi-weekly actual transmittance values for a base elevation from a file (see solrad.aml). Both values are generated by a program, which analyses SAMSON climate data (see sum_tran.for). Furthermore, the AML is enabled to run in batch mode so it allows the calculation of various time periods in batch mode (see solradb.aml)

    Diffuse solar radiation is calculated using the diffuse.aml routine (Kumar et al., 1997). The program was equally adjusted to run in batch mode (see diffuseb.aml). The calculations have been performed successfully.
     

    4.3. Development of a program to summarize SAMSON solar radiation data:

    A program was developed to analyze hourly data as provided by the SAMSON network on solar radiation and other meteorological variables (sum_srad.for). The program aims at summarizing the hourly data into daily sums, averages, and min/max statistics. The daily values are further summarized into monthly and yearly statistics. This allowed testing and improving the calculation of appropriate solar radiation values. Extensive tests on a series of SAMSON stations data demonstrated that it is unreliable to base the calculation of atmospheric transmittance on the diurnal temperature range (as proposed by Bristow and Campbell, 1984 ) and that environmental lapse rates that are used in MT-CLIM (Hungerford et al., 1989) are too low in our areas. Also, the tests have shown that elevational lapse rates for transmittance vary seasonally.

    The program then was adjusted to calculate daily, and bi-weeklysummary statistics for actual transmittance, clear-sky transmittance (defined as the bi-weekly maximum), as well as a series of other climatic variables. Some errors in the calculation of transmittance in the first program were corrected, but the analysis of the other climate variables (besids radiation) were dropped (see sum_tran.for). The elevational lapse rate is calculated as a function of several stations of the SAMSON network in the same climate zone.
     

    4.5. Etp, SWB, and SW models

    ArcInfo-AMLs have been developed and only need to be adjusted to the specific needs of this project. They are ready to use (see manuals).
     

    4.6. Soil information database

    For the Camp Williams site, the USDA soil survey map is available in digital form. It is labeled with the correct soil map units, which allows soil parameters to be appended to the spatial soils database. The necessary information includes (1) soil moisture holding capacity (mm water / m soil), (2) soil depth (m), and permeability of the soils (mm water / minute). The spatial database for the Camp Williams Soils is generated, the USDA Soil Survey data can now be downloaded and appended.A program was coded, which allows to integrate the complexe information about soil layers of soil map units as provided on the web (see soilprop.for). The program is very limited in its use, but serves the need of this project.
     
     
     

    5. Outlook:


    The following steps have to be performed next:


     
     

    6. References:

    Bristow, K.L. and Campbell, G.S. 1984. On the relationship between incoming solar radiation and daily maximum and minimum temperature. Agricultural and Forest Meteorology 31: 159-166.

    Caprio, J.M., 1974. The Solar Thermal Unit Concept in Problems Related to Plant Development and Potential Evapotranspiration. In: H. Lieth (Editor), Phenology and Seasonality Modeling. Ecological Studies. Springer Verlag, New York, pp. 353-364.

    Daly, C., Neilson, R.P. and Phillips, D.L. 1994. A Statistical-Topographic Model for Mapping Climatological Precipitation over Mountainous Terrain. Journal of Applied Meteorology 33: 140-158.

    Dubayah, R. and Rich, P.M., 1995. Topographic solar radiation models for GIS. International Journal for Geographical Information Systems, 9(4): 405-419.

    Gholz, H.L. 1982. Environmental limits on aboveground net primary production, leaf area, and biomass in vegetation zones of the Pacific Northwest. Ecology 63: 469-481.

    Grier, C.C. and Running, S.W., 1977. Leaf area of mature northwestern coniferous forests: relation to site water balance. Ecology, 58(4): 893-899.

    Hungerford, R.D., Nemani, R.R, Running, S.W. and Coughlan, J.C. 1989. MTCLIM: A mountain microclimate simulation model. USDA Forest Service, Research paper INT-414, Intermountain Research Station, Ogden.

    Jensen, M.E. and Haise, H.R., 1963. Estimating evapotranspiration from solar radiation. J. Irrig. Drainage Div. ASCE, 89: 15-41.

    Kumar, L., Skidmore, A.K. and Knowles, E., 1997. Modelling topographic variation in solar radiation in a GIS environment. International Journal for Geographical Information Science, 11(5): 475-497.

    Nemani, R.R. and Running, S.W. 1989. Testing a theoretical climate-soil-leaf area hydrologic equilibrium of forests using satellite data and ecosystem simulation. Agricultural and Forest Meteorology 44(3-4): 245-260.

    Rich, P.M., Hetrick, W.A. and Savings, S.C., 1995. Modelling topographical influences on solar radiation: manual for the SOLARFLUX model. LA-12989-M, Los Alamos National Laboratories, Los Alamos.

    Roberts, D.W., Fisher, R.F., Long, J.M. and Jack, S.N., 1993. The Leaf Area Allocation Model: Simulation of Rocky Mountain Forest Dynamics and Climate Change. Final Report for EPA Cooperative Agreement #817539.

    Running, S.W. 1994. Testing FOREST-BGC ecosystem process simulations across a climatic gradient in Oregon. Ecological Applications 4(2): 238-247.

    Running, S.W. and Gower, S.T. 1991. FOREST-BGC, A general model of forest ecosystem processes for regional applications: II. Dynamic carbon allocation and nitrogen budgets. Tree Physiology 9: 147-160.

    Thornton, P.E., Running, S.W. and White, M.A. 1997. Generating surfaces of daily meteorological variables over large regions of complex terrain. Journal of Hydrology 190(3-4): 214-251.

    Turc, L., 1963. Evaluation des besoins en eau d'irrigation, évapotranspiration potentielle, formulation simplifié et mise à jour. Ann. Agron., 12: 13-49.
     
     

    [top] [back] [home]
     
     
     

     Last Updated on 10/07/2003
    By Niklaus E. Zimmermann