Satellite Irrigation Management Support With the Terrestrial ...

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 5, NO. 6, DECEMBER 2012

1709

Satellite Irrigation Management Support With the Terrestrial Observation and Prediction System: A Framework for Integration of Satellite and Surface Observations to Support Improvements in Agricultural Water Resource Management Forrest S. Melton, Lee F. Johnson, Christopher P. Lund, Lars L. Pierce, Andrew R. Michaelis, Samuel H. Hiatt, Alberto Guzman, Diganta D. Adhikari, Adam J. Purdy, Carolyn Rosevelt, Petr Votava, Thomas J. Trout, Bekele Temesgen, Kent Frame, Edwin J. Sheffner, and Ramakrishna R. Nemani

Abstract—In California and other regions vulnerable to water shortages, satellite-derived estimates of key hydrologic fluxes can support agricultural producers and water managers in maximizing the benefits of available water supplies. The Satellite Irrigation Management Support (SIMS) project combines NASA’s Terrestrial Observation and Prediction System (TOPS), Landsat and MODIS satellite imagery, and surface sensor networks to map indicators of crop irrigation demand and develop information products to support irrigation management and other water use decisions. TOPS-SIMS provides the computing and data processing systems required to support automated, near real-time integration of observations from satellite and surface sensor networks, and generates data and information in formats that are convenient for agricultural producers, water managers, and other end users. Using the TOPS modeling framework to integrate data from multiple sensor networks in near real-time, SIMS currently maps crop fractional cover, basal crop coefficients, and basal crop evapotranspiration. Map products are generated at 30 m resolution on a daily basis over approximately 4 million ha of California farmland. TOPS-SIMS is a fully operational prototype, and a publicly available beta-version of the web interface is being pilot tested by farmers, irrigation consultants, and water managers in Manuscript received November 01, 2011; revised March 09, 2012; accepted April 10, 2012. Date of publication December 11, 2012; date of current version December 28, 2012. TOPS-SIMS development was funded by the NASA Applied Sciences Program under Award NNX10AE48A to California State University, Monterey Bay (CSUMB). F. S. Melton, L. F. Johnson, C. P. Lund, L. L. Pierce, A. R. Michaelis, S. H. Hiatt, A. Guzman, A. J. Purdy, C. Rosevelt, and P. Votava are with the California State University, Monterey Bay, Seaside, CA 93955 USA (e-mail: [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected]). F. S. Melton, L. F. Johnson, C. P. Lund, A. R. Michaelis, S. H. Hiatt, A. Guzman, P. Votava, E. J. Sheffner, and R. R. Nemani are with the NASA Ames Research Center, Moffett Field, CA 94035 USA (e-mail: [email protected], [email protected]). D. D. Adhikari is with the Center for Irrigation Technology, California State University Fresno, Fresno, CA 93740 USA (e-mail: [email protected]). T. J. Trout is with the USDA Agricultural Research Service, Fort Collins, CO 80526 USA (e-mail: [email protected]). B. Temesgen and K. Frame are with the California Department of Water Resources, Sacramento, CA 95814 USA (e-mail: [email protected], [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/JSTARS.2012.2214474

California. Data products are distributed via dynamic web services, which support both visual mapping and time-series queries, to allow users to obtain information on spatial and temporal patterns in crop canopy development and water requirements. TOPS-SIMS is an application framework that demonstrates the value of integrating multi-disciplinary Earth observation systems to provide benefits for water resource management. Index Terms—Agriculture, irrigation, remote sensing, soil moisture, water resources, wireless sensor networks, web services.

I

I. INTRODUCTION

N the western US, substantial year-to-year variability in weather conditions complicates planning for water managers and agricultural producers, and makes it difficult to apply standard rule-of-thumb approaches to agricultural water management. In California, the recent three year drought, from 2007–2009, underscored the vulnerability of the state’s agricultural industry to variability in water supply, with hundreds of millions in economic losses and thousands of agricultural jobs lost [1]. Concurrent with the drought, groundwater levels in the Central Valley declined substantially [2] as growers pumped water to partially offset reductions in surface water deliveries. Inter-annual variability in water supply is common throughout the western U.S., where growing urban populations, increases in water allocations for endangered species, and climate change—with predicted shifts in precipitation and reductions in winter snowpack in the Sierra and Rocky Mountains [3]–[5]— will continue to constrain water supplies for agricultural users. Because agriculture is typically the largest user of water in the West, maximizing the benefits of available agricultural water supplies is central to the implementation of drought mitigation and climate adaptation strategies in western states. Low cost adaptation strategies for agricultural water users include the use of irrigation scheduling technologies that incorporate information on weather conditions and crop growth stage. Satellite measurements of water resources and hydrologic fluxes have great potential for supporting agricultural producers and water managers working to optimize agricultural water supply and use. When combined with biophysical models, satellite observations can be applied to scale point observations from

1939-1404/$31.00 © 2012 IEEE

1710


TABLE I SUMMARY OF KEY TERMS

surface observation networks across larger geographic areas. Evapotranspiration (ET) and soil moisture data are particularly relevant to both operational water management and long-term planning. Surface measurements of these parameters are often difficult and expensive to collect, and generally available only from sparsely distributed observation networks. Recent advances in modeling and measurement of ET using satellite observations now enable moderate resolution (30–1000 m), spatially continuous estimates of ET over large areas in near real-time [6]–[9]. Integration of this information into current water management practices and operational models, however, requires the development of frameworks for synthesizing data from multiple sensor networks to drive models for estimating evapotranspiration, soil moisture, and other key hydrologic reservoirs and fluxes. New approaches are also required for delivery of satellite-derived data and information products to water managers and agricultural producers. One water management challenge where integration of ET models and sensor network observations has immediate potential utility is in supporting agricultural producers and water managers working to optimize irrigation management by accounting for current weather and crop conditions when

AND

ABBREVIATIONS

making irrigation management decisions. In scheduling irrigation, growers must account for multiple factors, including soil type, salinity and leaching fraction, weather conditions and crop growth stage. In water limited regions, it is particularly important for growers to maximize on-farm water use efficiency by accounting for weather conditions in estimating crop water consumption and associated irrigation demand. A common method of irrigation scheduling involves estimation of crop water use by multiplying weather-based estimates of reference evapotranspiration by crop coefficients for a particular crop. (A summary of key terms and definitions related to agricultural irrigation management is included in Table I.) is defined as the evapotranspiration from a reference crop, typically a well-watered, actively growing grass or alfalfa of uniform height. In the western U.S., agricultural weather networks such as the California Irrigation Management Information System (CIMIS), AgriMet, and the Arizona Meteorological Network (AZMET) provide hourly information on weather conditions and . Within California, CIMIS [10] currently operates 139 stations across the state and provides data to growers and irrigation consultants on and weather conditions. Spatial CIMIS daily data derived using the

MELTON et al.: SATELLITE IRRIGATION MANAGEMENT SUPPORT WITH THE TERRESTRIAL OBSERVATION AND PREDICTION SYSTEM

methods developed by Hart et al. [11] are available from 2003 to present at 2 km spatial resolution statewide. The method uses cloud cover estimates derived from the GOES weather satellites to estimate the daily solar insolation for each pixel, and thus represents an initial integration of satellite and surface observations within CIMIS. Guideline crop coefficients have been compiled for several crops and are available from sources such as the Food and Agriculture Organization Irrigation and Drainage Paper 56, commonly known as FAO-56 [12], the University of California’s Basic Irrigation Scheduler, the California Department of Water Resources’ Consumptive Use Program, and the Wateright system developed by California State University, Fresno. is a dimensionless value representing typical crop evapotranspiration for a given crop canopy relative to , and it captures the effect of crop growth stage on evaporative demand. integrates the effects of both crop transpiration and soil evaporation. Many growers use a dual-crop coefficient approach, which separates the evaporation and transpiration components of ET. The basal crop coefficient represents conditions for an unstressed crop on a dry soil surface, and a separate coefficient is used to calculate soil evaporation if required. can be used to calculate basal crop ET as a measure of plant transpiration for a non-stressed crop plus a small diffusive soil evaporation component. As such, also provides a measure of potential crop water use, or biological water demand for well-watered crops. While useful for seasonal planning, one limitation of this standard approach to operational irrigation management is that it requires growers to derive crop coefficients from look-up tables and then correct them for site-specific conditions. In addition, many crop coefficients were developed decades ago and may not account for current farming practices or local and interannual variability in weather conditions and crop development rates. Previous research conducted in California quantified the observed benefits of using information on ET in irrigation scheduling. This research compared applied irrigation and yields reported by growers on 55 farms before and after use of data from CIMIS. Results from the study reported an average increase in yields of 8%, and a reduction in total applied irrigation of 13% [13], [14]. While it is difficult to fully control for the effect of year-to-year variability in weather on crop yields in these types of studies, this result suggests that during key phases in crop development, some growers may be irrigating to the point that the root zone becomes saturated, reducing oxygen availability in the root zone and modestly suppressing crop growth. Across the 13 different crop types evaluated, including tree and vine crops, vegetable crops, and field crops, growers reported total economic benefits from increased yields and reductions in applied water ranging from $110 to $1,700 per hectare. Due in part to these benefits, use of CIMIS data in California has grown from about 450 primary registered CIMIS data users in 1987, to nearly 40,000 today. In the year 2010 alone, for example, reports were generated from the CIMIS database. However, the most recent farm and ranch irrigation survey conducted by the USDA indicated that only 12% of growers in California utilize reports on daily crop-water ET in scheduling irrigation [15], due in part to the barriers described above. Devel-

1711

opment of an automated system for mapping crop coefficients and delivery of data to users would enhance the use of ET data in irrigation management. The California Department of Water Resources (CDWR) identified statewide mapping of crop coefficients in near-real-time as one of the high priority, short term needs for CIMIS [16]. In addition to , observations of other parameters related to agricultural water management are increasingly available from surface sensor networks and satellites. Soil moisture information is available for a limited number of sites from the USDA Soil Climate Analysis Network (SCAN). As the cost of soil moisture sensors has declined, an increasing number of specialty crop producers are also deploying soil moisture sensor networks in their operations. Satellite and aircraft remote sensing observations of crop canopies over agricultural regions are used by the USDA National Agricultural Statistics Service (NASS) to map crop types and fallowed acreage on an annual, post-season basis, and increasingly by individual growers to identify and map variations in crop canopy conditions that may be indicative of pest infestations, variations in soil texture, or irregularities in irrigation systems. Seasonal estimates of evapotranspiration from satellite-driven energy balance models such as the Surface Energy Balance Algorithm for Land (SEBAL) [17] and the Mapping EvapoTRanspiration with Internalized Calibration (METRIC) [18] model have also been used by water managers for water rights accounting to resolve disputes over water resources [19]. Integration of these data sources into current management practices and operational models presents a number of challenges, especially for applications that require information in near real-time. In addition to typical issues associated with processing large volumes of data from heterogeneous sources, data from the Thematic Mapper (TM) and Enhanced Thematic Mapper sensors onboard the Landsat satellites must be atmospherically corrected to minimize artificial scene-to-scene variability resulting from the effects of aerosols, ozone, and water vapor. A hardware failure on the Landsat 7 instrument produces regularly spaced data gaps that must be addressed. Differences in the spatial scale of different satellite sensors must be considered when fusing data from sensors such as Landsat TM and (30 m–120 m) and the MODerate resolution Imaging Spectroradiometer (MODIS) (250 m–1 km) to reduce data gaps due to clouds and improve data reliability. Data from surface sensor networks are valuable for calibrating and scaling satellite-driven models, but they must be retrieved and transformed into formats that can be ingested by various models. At present, a number of satellite-driven energy balance models (including SEBAL and METRIC) for estimation of ET also require manual intervention to select “cold” and “hot” anchor pixels that serve to constrain the calculation of ET, making it difficult to fully automate operation of these models, though work by Kjaersgaard et al. [20] shows progress towards resolving this problem. As a result of these issues, the full potential of satellite data for supporting water managers and agricultural producers has yet to be fully realized. The NASA Terrestrial Observation and Prediction System (TOPS) [21] provides a modeling and computing framework for integrating satellite and surface observations in near real-time

1712


at the field level. By accounting for the irrigation application rate, distribution uniformity, and any corrections for water replacement targets (e.g., intentional deficit irrigation or leaching requirement), users can translate TOPS-SIMS output into an irrigation management strategy. To evaluate the utility of the system for commercial farming operations, TOPS-SIMS estimates are compared against ET data collected by surface renewal stations [22] and information from wireless soil moisture sensor networks deployed in operational agricultural fields across California in cooperation with growers. The TOPS-SIMS architecture facilitates forecasting of potential consumptive use and associated irrigation demand with lead times of up to 1 week through integration of forecast data from the NOAA National Weather Service (NWS) Forecasted Reference Evapotranspiration system (FRET). This capability is particularly relevant for future extension of TOPS-SIMS to support both irrigation scheduling and water delivery planning. Finally, the TOPS-SIMS architecture is designed to support the future addition of other satellite-driven models for estimation of ET, soil moisture, and crop stress, all of which further the goal of improving water management and making irrigation scheduling more practical, convenient, and accurate. Fig. 1. Overview of the TOPS-SIMS architecture with primary data inputs and outputs.

to address many of the challenges described above. In the following sections, we describe an application of the TOPS framework to develop a system for near real-time mapping of crop canopy conditions and associated crop irrigation requirements at the resolution of individual fields. To support continued improvements in management of agricultural water supplies, the data processing system is designed to provide growers and water managers with frequent, accurate, and easily available information that includes key indicators of crop canopy development and crop water use requirements. II. A FRAMEWORK FOR SATELLITE IRRIGATION MANAGEMENT SUPPORT The TOPS Satellite Irrigation Management Support (TOPSSIMS) framework integrates satellite observations from Landsat and MODIS with meteorological observations from CIMIS and ancillary data on crop type and site specific conditions (Fig. 1). The system employs a modular architecture to facilitate support for a wide range of models and data. The initial implementation provides a capability for mapping of fractional cover , associated basal crop coefficients , and evapotranspiration over 3.7 million ha of farmland in California’s Central Valley. and are mapped every eight days, and maps are produced on a daily basis at spatial resolutions that are useful for irrigation management at the field level (30 m). Automated atmospheric correction and gap-filling algorithms are optimized for agricultural areas to provide a robust and reliable data stream. Information from TOPS-SIMS is distributed to water managers and agricultural producers via a browser-based irrigation management decision support system. Users may combine the supplied or with formal or intuitive estimates of soil evaporation to derive total water consumption over the recent period

III. METHODS TOPS, a NASA modeling and data assimilation framework developed to monitor and forecast environmental conditions, provides a range of capabilities for integrating observations from satellites and surface sensor networks via models including crop models, biogeochemical and carbon cycle models, ecological models, and numerical weather and climate models [21]. The TOPS framework has also been adapted to run on high-end computing resources in the NASA Advanced Supercomputing (NAS) facility as part of the NASA Earth Exchange (NEX) [23]. NEX combines supercomputing resources, Earth system modeling, remote sensing data from NASA and other agencies, and a scientific social networking platform to deliver a complete work environment in which users can analyze large Earth science data sets, run modeling codes, collaborate on new or existing projects, and share results within and among communities. NEX provides direct access to the full archive of Landsat and MODIS data for the western U.S., as well as on-demand access to thousands of processing cores, facilitating rapid processing of both historic data archives and incoming satellite data. Leveraging these capabilities, TOPS-SIMS ingests Landsat-5 TM, Landsat 7 , and MODIS satellite imagery for California’s Central Valley and other agricultural regions. Landsat provides the spatial resolution necessary to produce information at the scale of individual fields. The daily temporal resolution of MODIS provides a gap-filling capability to ensure data availability. Methods employed by the remote sensing community for satellite mapping of ET at field scales (30 m) rely primarily on energy balance models, e.g., [17], [18], [24], or reflectancebased mapping of crop coefficients, e.g., [25]–[28]; see also review by Gowda et al. [29]. Of these approaches, strategies for full automation at the scale of individual fields have only been developed for mapping of crop coefficients from surface reflectance, and thus the initial suite of algorithms implemented in the TOPS-SIMS framework relies on this approach. In the


Fig. 2. Past studies conducted by USDA in collaboration with NASA [35], [36] provide the basis for linking NDVI to fractional cover using relationships that (a). are robust across different crop types and canopy architectures In (b), use of a generalized equation (after methods of Allen and Pereira [40]) for converting fractional cover to FAO crop coefficients provides a robust approach values for annual crops . When the crop type for mapping at a particular location is specified or can be identified with confidence, crop specific equations may also be applied.

future, it may be possible to implement additional models, including energy balance models and models for downscaling of soil moisture estimates from satellite missions such as the Soil Moisture Active Passive (SMAP) mission. Each Landsat scene is atmospherically corrected using software developed under the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) project [30]. LEDAPS incorporates the 6S atmospheric radiative transfer modeling approach of Vermote et al. [31], as included in the standard MODIS data processing chain. Landsat TM and data are tiled onto a common grid to form an 8-day composite. This facilitates use of overlapping portions of each scene to increase the frequency of observations and reduce data gaps due to cloud cover. This compositing approach also reduces the effects of the scan line correction error, as only the central portion of the Landsat 7 scenes are used, where the gaps are narrower. The next step is to calculate the normalized difference vegetation index (NDVI) [32] from the composited Landsat scenes. NDVI is an index calculated from the red and near infrared wavelengths and provides a measure of photosynthetic capacity. At this stage, a suite of gap-filling algorithms are employed to ensure spatially continuous data over agricultural regions. Current algorithms include the use of moving window averages and linear interpolation to fill small data gaps over agricultural fields. Field boundaries available from many of the County Agricultural Commissioners’ offices in California facilitate

1713

identification of MODIS 250 m NDVI pixels that are fully contained within agricultural field boundaries, and thus enable use of MODIS data to fill the majority of the remaining data gaps with reasonable accuracy. Use of the STAR-FM algorithm [33], as implemented within TOPS, can be employed to address any remaining gaps through fusion of Landsat and MODIS by disaggregating MODIS data to the Landsat scale. Prior to further processing, non-agricultural areas are masked out using the USDA NASS Cropland Data Layer (CDL)1 [34] and field polygons from local County Agricultural Commissioner offices, when available. The CDL is also used to spatially differentiate annual crops from perennial orchards and vineyards. Following the compositing and gap-filling procedures, NDVI data are transformed to via empirical relationships developed by USDA and NASA. Johnson and Trout [35] and Trout et al. [36] collected field measurements of across multiple crop types in the Central Valley. These data were compared with satellite observations collected on the same day, and revealed a robust relationship between NDVI and (Fig. 2(a)). This approach is consistent with previous studies showing that various spectral vegetation indices, calculated from visible and near-infrared (NIR) reflectance data, are linearly related to canopy light interception [37]–[39]. To convert to , TOPS-SIMS uses different approaches . For retfor retrospective versus near real-time mapping of rospective mapping of annual crops or for fields where information on the current crop type is available from the grower, is converted to based on a physical description of the crop canopy [40]. A density coefficient , based on and estimated plant height, is calculated first to link increases in the is then crop coefficient with increasing vegetation amount. used to interpolate between the minimum and maximum for a given crop type, as listed in FAO-56. Using methods deconscribed by Allen and Pereira [40], a crop specific version can be applied retrospectively to a given field based on the CDL crop classification or information provided by the grower or farm manager. The CDL is not publicly available until after December 31 of each year, however, presenting a challenge using this approach. While for near real-time mapping of the crop type for perennial fields (vineyards, orchards) can be reasonably determined from prior-year maps, the spatial distribution of annuals, such as lettuce, tomato, or cotton, is generally unknown within-season due to the complex and diverse nature of the agricultural cropping patterns throughout much of Calirelationship representing a best-fit fornia. A generalized to crop specific relationships (Fig. 2(b)) is thus applied for near real-time mapping of fields deemed by CDL to contain annual mean estimation crops. Use of this approach introduces uncertainty ranging from 3–14% across the major crop categories (vegetables, roots/tubers, legumes, fibers, oils, cereals). For perennial crops, results from multi-year studies recently performed on large weighing lysimeters at the University of Calconifornia Kearney Agricultural Center are used for version [41], [42]. Both studies reported a strong relationship between mid-day canopy light interception, which is closely for both related to , and the crop coefficient. Finally, 1[Online]. Available: http://www.nass.usda.gov/research/Cropland/ SARS1a.htm (last accessed 07May2012)

1714


Fig. 3. Examples of comparisons between profiles from TOPS-SIMS using generalized (solid blue line) and crop-specific (dashed blue line) relationships versus FAO-56 [12] (solid green line) for various crops during 2011. The comparisons highlight the ability of TOPS-SIMS to modify the idealized FAO-56 for observed crop cover and phenology as related to site-specific management practices and local weather conditions.

annuals and perennials is calculated as the product of and . For California, is retrieved by TOPS-SIMS via FTP (file transfer protocol) from the standard Spatial CIMIS 2 km daily statewide grids. TOPS-SIMS simplifies the process of deriving crop coefficients for estimating using an automated data integration and processing system, fully independent of contemporaneous crop type classification, and provides a framework for assimilating other sensor network observations and models to provide information on crop canopy conditions and ET. Crop coefficient values from the fully automated TOPS-SIMS compare well with crop coefficients derived from FAO-56 planning guidelines during initial growth and at near full cover, but can deviate during vegetative growth and maturation. Comparisons for individual fields highlight differences due to agronomic practices and crop response to site-specific weather conditions (Fig. 3). For example, the seasonal profiles for tomatoes and cotton closely match the FAO-56 profiles when using a crop-specific formula to convert to , while lettuce and watermelon were better characterized by the generalized conversion. Corn and wheat observations reveal different development rates and overall phenological progression than those of FAO. Johnson and Melton [43] also compared TOPS-SIMS estimates against values

from SEBAL, which employs a more complex energy-balance approach. SEBAL has been widely used to model ET over agricultural regions, but requires a highly trained analyst to manually perform the required calibrations and operate the model. TOPS-SIMS values compared favorably with SEBAL estimates for unstressed crops for representative San Joaquin Valley crop types, with mean differences ranging from 0.4 to 0.85 mm/day (Fig. 4). IV. INTEGRATING SATELLITE AND SURFACE SENSOR NETWORKS WITH ET MODELS Automated integration of satellite observations of crop canopy development with surface observations of is an important first step in developing a fully integrated sensor network or “system of systems” capable of providing estimates of ET and soil moisture derived from multiple data sources to water managers and agricultural producers. Integration of soil moisture observations from soil moisture sensor networks installed in agricultural fields is an important additional advance, as soil moisture observations can provide operational users with an independent check on a water balance derived from satellite estimates and field-specific irrigation and weather records. Integration of soil moisture data also provides a mechanism for future automation of model calibration to


1715

Fig. 4. Comparison of daily water use estimates (mm) for unstressed San Joaquin Valley fields generated by two satellite-driven models (from Johnson and Melton generated using the SEBAL surface energy balance approach. Basal evapotranspiration generated by TOPS-SIMS. [43]). Actual evapotranspiration Mean absolute error (MAE) in mm/day was calculated as a measure of model agreement, as shown. All dates 2009.

account for intentional deficit irrigation by the grower. It is cost prohibitive to collect soil moisture observations across an entire field, but selective use of soil moisture sensors can increase confidence in satellite-derived estimates and facilitate use of satellite data to scale information derived from soil moisture observations over larger areas. The cost of soil moisture sensors and wireless data loggers has declined substantially over the past decade, and there are now a number of commercially available options for growers interested in deploying soil moisture sensor networks. In addition, new cost effective techniques have been developed for measuring actual ET in agricultural environments characterized by fields on the order of 10–40 hectares. While techniques such as the surface renewal approach [22] are still too expensive and complex for use by commercial farms growing specialty crops, field instrumentation has now become more cost effective, such that multiple sites can be monitored by water management agencies or small research teams providing important reference and validation data. To assist in understanding relationships between satellite estimates of and surface observations of and soil moisture, we deployed soil moisture sensor webs on 12 commercial farms in the Central and Salinas Valleys, in research fields located at California State University Fresno in Fresno, CA, and at the University of California West Side Research and Extension Center in Five Points, CA. Sensors deployed included capacitance probes for volumentric soil water content measure-

ment, soil water potential sensors, passive capillary lysimeters for measuring deep drainage, in-line flow meters for measuring total applied irrigation, aboveground pheno-cams (which record digital images of the field once per day), and meteorological stations. In collaboration with CDWR, we also deployed and operated above-ground surface renewal stations at seven sites in a range of crops to provide measurements of each component of the surface energy balance, facilitating calculation of for each site. In surface renewal analysis, ET is calculated as the residual of a surface energy balance equation using measured net radiation, ground heat flux, and high-frequency measurements of air temperature [22]. These temperature traces exhibit low-frequency ramp-like structures which are used to calculate sensible heat flux. Surface renewal makes no assumptions regarding atmospheric stability, and is here combined with sonic anemometer data to account for unequal heating of air parcels across heterogeneous landscapes [44]. Integration of data from these sources has clear value for understanding spatial and temporal patterns in ET and soil moisture. Integration of satellite derived data with observations of volumetric soil water content and soil water potential has particular value for optimization of irrigation scheduling. Fig. 5(a) provides an example for a sub-surface drip irrigated cotton field located near Huron, CA. The satellite data provides information on the day-to-day crop water use under well-watered conditions, while the soil moisture sensors assist in identifying the onset of water stress and magnitude of deficit irri-

1716


Fig. 5. To illustrate the benefits of integrating satellite estimates of with soil moisture measurements from surface sensor networks, measures of ET and (green line), combined total irrigation soil moisture from different sources are plotted for a cotton field near Five Points, CA. Cumulative TOPS-SIMS and precipitation (light-blue dashed line), net water balance (dark-blue dotted line), and observed volumetric soil water content (VWC, brown line) are plotted begins to exceed daily precipitation and applied irrigation as indicated by the negative slope of the net water balance line. together in (a). On June 8, daily By July 4 the cumulative water balance turns negative (indicating potential onset of plant water stress), and deficit irrigation is sustained through August. In (b), of 0.85 to account for soil water limitations on plant transpiration from July 4 onward (solid green line) substantially incorporation of a crop stress coefficient estimates from TOPS-SIMS (dashed green line) and measurements from a surface renewal station located in the field improves agreement between (solid dark-blue line).

gation. The net water balance, or change in total estimated soil water content , is calculated as:

where is precipitation in mm measured by the on-site weather station; is irrigation in mm, calculated as the average flow measured in two drip laterals and divided by effective irrigated

area; D is drainage below the rooting zone measured with a Gee capillary lysimeter installed at a depth just below the rooting zone (120 cm); and is the basal crop ET in mm. The daily measured volumetric soil water content (VWC) is calculated as the average of the volumetric water content measured using capacitance probes (Decagon 10 HS) installed at three depths (0–10 cm, 35–45 cm, and 70–80 cm) at each of eight locations throughout the field. is the measured ET from the surface


1717

renewal station located in the center of the field. As shown in Fig. 5(b), and agree well during the early stages of crop development when the crop is well-watered. After June 8, 2011, daily exceeds total applied irrigation and after July 4, the cumulative water balance turns negative. This is reflected in the observed VWC which declines from June 8 onwards. The sustained negative net water balance also reflects intentional deficit irrigation by the grower, and this deficit is captured by the difference between the surface renewal measurements and the satellite-based estimates. Plants respond to water stress by reducing transpiration, and under the FAO-56 approach, a crop water stress coefficient based on soil water content is used to account for this reduced transpiration in calculating [12]. The addition of a coefficient of 0.85 to account for the deficit irrigation results in close agreement between the satellite-derived and surface renewal estimates (Fig. 5(b)). This example illustrates the potential benefits of using surface observations of soil water content or soil water potential measurements in coordination with estimates of that integrate satellite and surface observations. The satellite-derived information can assist with tracking the daily irrigation demand, while soil moisture sensor networks provide an important check to ensure that the timing and magnitude of irrigation deficits are in-line with targets set by the grower. V. DATA DISTRIBUTION VIA WEB INTERFACE SERVICES

AND

WEB

While TOPS-SIMS simplifies the process of integrating satellite and surface observations to map and values, distribution of data and information via intuitive, easy to use interfaces is critical to facilitating use of the information by water managers and agricultural producers who may not be familiar with the specialized data formats used by the remote sensing science community. A web-based user interface providing access to visualizations of TOPS-SIMS information eliminates the barriers to data access, as only a web browser is needed to view and query the information, and no knowledge of specialized data formats is necessary. The TOPS-SIMS web-based front-end relies on an open-source software stack that utilizes the Open Layers JavaScript library to display maps and execute queries on map data from within web browsers (Fig. 6(a)). To facilitate geographic searches by users and provide a spatial context for the TOPS-SIMS information products, the interface combines TOPS-SIMS overlays with a selection of open-source and commercial base map layers, including maps of streets, land use, and aerial imagery. NDVI, , and overlays are updated on nominal 8-day intervals; is updated daily. In addition to providing views of TOPS-SIMS data at the native 30 m resolution, the web interface also allows users to retrieve quantitative information from any variable by specifying a point or polygon and requesting the most current values or a time-series summary (Fig. 6(b)). The TOPS-SIMS web interface relies on data services provided by an open-source software stack running a Python-based web server. Tools from the Geographic Data Abstraction Library (GDAL) generate the tiles used in the web maps. The server responds to map queries by retrieving data resting on

Fig. 6. The TOPS-SIMS web interface provides capabilities for visualization and data access, including support for geographic queries and selection of data (mm/day) for 3.7 million for visualization by date. The map in (a) shows hectares of farmland in the Central and Coastal Valleys on 02-Aug-2011. Retrospective time-series for TOPS-SIMS output can be generated, as shown by graph for 2011 for a location near Huron, CA (b). Time-series data can the be downloaded from the interface in CSV format directly into a spreadsheet for further analysis.

an OPeNDAP web service (Open-source Project for a Network Data Access Protocol; http://opendap.org) and generates time-series charts from the resulting data using the Matplotlib graphics library. OPeNDAP is a software framework that facilitates remote access to local data regardless of local storage format, and thus simplifies the process of making scientific data sets available for other applications via web services. A working beta-version of the interface and associated web data services are available at http://ecocast.arc.nasa.gov/sims/, and are currently being evaluated by commercial growers, irrigation consultants, and water managers in California through a series of ongoing pilot studies.


1718

VI. EXTENDING THE TOPS-SIMS FRAMEWORK The current TOPS-SIMS and mapping capabilities demonstrate the potential utility of a modeling framework to integrate observations from satellite and surface sensor networks to provide new data and information services to the agricultural and water resource management user community. However, the TOPS-SIMS modular architecture also allows for a number of extensions that can be rapidly implemented to expand the data and information services. Key short term objectives include: i) linking TOPS-SIMS to forecasts to provide forecasts of irrigation demand at various lead times; ii) use of web services and a model web approach [45] to integrate TOPS-SIMS outputs with operational and planning models utilized by water managers; iii) incorporation of site specific information provided by growers to include generation of customized reports; iv) expansion of inputs to include observations from a constellation of moderate resolution satellite sensors to increase observation frequency and improve the reliability and accuracy of data and information products from TOPS-SIMS; and v) incorporation of additional publicly available models for estimation of ET, including satellite-driven energy balance models for estimation of ET. At present, managers of agricultural water supplies generally rely on historic climate data and static estimates of irrigation demand when making decisions about the volume of water to release from reservoirs. For agricultural water delivery systems such as the Central Valley Project (CVP) and the California State Water Project (SWP), which together provide water for 3.75 million acres of California farmland, water managers must rely on this historical data to time water releases up to a week or more in advance to meet the needs of growers. While water managers must also consider a range of other factors, including environmental constraints and hydropower production, new sources of data on forecasted irrigation demand could support improvements in the ability to anticipate periods of particularly high demand in response to high temperatures or sustained winds, while conserving water during cooler periods with lower irrigation demand. The recent development of two experimental forecasting systems provides the possibility of using forecasts of short-term water consumption in place of estimates derived from historic climate data and maps of cropping patterns. The NOAA National Weather Service (NWS) Forecasted Reference ET (FRET) system2 forecasts with a lead time of up to seven days. The FRET forecasts for California are driven by outputs from the Weather Research and Forecasting model (WRF) [46], and utilize WRF forecast fields to parameterize the Penman-Monteith equation for as adopted by the Environmental Water Resources Institute, American Society of Civil Engineers [47]. Experimental 8-day forecasts and seasonal outlooks are also being developed by the National Center for Atmospheric Research (NCAR) and the Bureau of Reclamation, and employ statistical forecasting methods to provide probabilistic outlooks3. Data from the NOAA 2[Online]. Available: http://www.wrh.noaa.gov/forecast/evap/FRET/FRET. php?wfo=sto (last accessed 07May2012) 3[Online].

Available: http://calpet.org/ (last accessed 07May2012)

NWS FRET system are already being ingested by TOPS-SIMS via a web service, and can be directly integrated with maps to provide estimates of weekly irrigation demand over California at the scale of individual fields. Pending the outcome of ongoing analyses to identify biases and quantify uncertainty in these forecasts relative to observed values from CIMIS, data from these forecast systems will be applied to produce and distribute forecasts to agricultural producers and water managers to assist in quantification of local, regional, and statewide crop water consumption and irrigation demand. Improved data on current and forecasted irrigation demand that incorporates satellite observations of crop conditions can also be incorporated into operational models utilized by the CDWR such as CALSIM [48], and planning models such as Cal-SIMETAW [49], [50]. Incorporating recent historic data into these models has potential utility for evaluating past decisions on water releases, identifying opportunities for future improvement in delivery of water to growers, and developing strategies for implementation of water markets during drought years. In cases where forecasted demand is significantly higher or lower than historic demand for a given period, water operations managers could adjust management practices and improve their ability to deliver water to growers when and where it is most needed. Satellite derived and maps are also potentially valuable as inputs into longer term planning models, and the development of a year record of observed crop canopy conditions and values could lead to improvements in models used for long term planning, such as Cal-SIMETAW. Because TOPS-SIMS is implemented on NEX and data products are already distributed via an OPeNDAP web service, integration with other models is straight-forward, as data can be requested via existing web services, and data not already available in the NEX archive can be rapidly generated by TOPS-SIMS. Incorporation of site-specific information from growers can assist in eliminating sources of uncertainty in estimating ET associated with crop type, the timing and amount of irrigation applied, and the irrigation system configuration. Research is ongoing to extend the TOPS-SIMS framework to allow growers to specify this information for individual fields, facilitating use of crop-specific relationships for mapping from satellite data, and allowing growers to specify correction factors to account for intentional deficit irrigation or a leaching fraction. Grower information on precedent irrigation schedule and irrigation system type would allow estimation of soil evaporation and conversion of to . Allowing growers to input information on irrigation system type and application rates would also allow TOPS-SIMS to translate into recommended irrigation system run times. TOPS-SIMS algorithms for deriving from NDVI currently rely on data from the Landsat TM, Landsat , and MODIS satellite sensors, but NDVI can also be derived from observations from multiple other satellite sensors currently collecting data in the red and near-infrared wavelengths at resolutions from 20 m to 60 m. Data from the SPOT-5 satellite instruments, the Indian Resource Satellite Advanced Wide Field Sensor (AWiFS), and the upcoming Landsat Data Continuity Mission (LDCM) and Sentinel-2 satellite mission offer the po-


tential for integration of observations into a “system of systems” to provide field-scale observations of crop canopy development multiple times per week, reducing data latency and data gaps resulting from cloud cover. While issues associated with inter-sensor calibrations must be fully resolved [51] and open data policies have not yet been finalized for all of these missions, TOPS-SIMS provides a framework for ingesting and compositing data from a constellation of moderate resolution (30–250 m) sensors to provide robust and reliable data and information services to agricultural producers and water managers. Finally, TOPS-SIMS is designed to support multiple models for estimation of ET from satellite observations, including energy balance models that incorporate information from the thermal bands on Earth observation satellites. The initial suite of algorithms implemented within TOPS-SIMS relies on the use of spectral information in the visible and near-infrared wavelengths to map and values since this approach can be fully automated. As research progresses on automation of the internal calibration procedures for energy balance models [20], TOPS-SIMS is designed to support integration of these models, and to facilitate use of a suite of publicly-available models to provide a robust capability for automated, frequent, field-scale mapping of ET over regions the size of the western U.S. VII. SUMMARY TOPS-SIMS employs a “system of systems” approach and applies the TOPS modeling framework to ingest observations from satellite and surface sensor networks to provide new data and information products to agricultural producers and water managers via easily accessible web interfaces and web services. The current framework supports near real-time mapping of indicators of crop canopy development and crop water consumption at field scales over 3.7 million ha of California farmland. Integration of satellite-derived estimates of basal crop evapotranspiration with observations of soil moisture from surface sensor networks offers promise for supporting agricultural producers and water managers working to optimize management of agricultural water resources. Use of the NEX computing architecture enables rapid processing of large volumes of satellite data and facilitates implementation of TOPS-SIMS over regions potentially as large as the western U.S. TOPS-SIMS is designed to integrate additional models and data services to support forecasting of crop irrigation requirements at weekly to seasonal lead times, and concurrent modeling of potential and actual crop evapotranspiration. Future integration of observations from a constellation of moderate resolution satellites would support further improvements in the frequency and long-term operational reliability of satellite-derived estimates of evapotranspiration and crop water requirements. ACKNOWLEDGMENT The energy-balance comparison data were generated by SEBAL North America, Inc., under a sub-contract from the University Corp. at Monterey Bay. The authors gratefully acknowledge Ian Harlan and Randall Holloway of CSUMB for assistance with deployment of field instrumentation, as well as the anonymous reviewers for their comments and suggestions to improve the manuscript.

1719

REFERENCES [1] R. Howitt, J. Medellin-Azuara, and D. MacEwan, “Measuring the employment impact of water reductions,” Univ. California, Davis, CA, 2009. [2] J. S. Famiglietti, M. Lo, S. L. Ho, J. Bethune, K. J. Anderson, T. H. Syed, S. C. Swenson, C. R. de Linage, and M. Rodell, “Satellites measure recent rates of groundwater depletion in California’s central valley,” Geophys. Res. Lett., vol. 38, p. L03403, 2011. [3] R. Seager, M. Ting, I. Held, Y. Kushnir, J. Lu, G. Vecchi, H. P. Huang, N. Harnik, A. Leetmaa, N. C. Lau, C. Li, J. Velez, and N. Naik, “Model projections of an imminent transition to a more arid climate in southwestern North America,” Science, vol. 316, pp. 1181–1184, 2007. [4] M. R. Shaw, L. Pendleton, D. Cameron, B. Morris, G. Bratman, D. Bachelet, K. Klausmeyer, J. MacKenzie, D. Conklin, J. Lenihan, E. Haunreiter, and C. Daly, “The Impact of Climate Change on California’s Ecosystem Services,” California Energy Commission, Sacramento, CA, CEC Report #CEC-500-2009-025-D, 2009. [5] J. Overpeck and B. Udall, “Dry times ahead,” Science, vol. 328, pp. 1642–1643, 2010. [6] A. Calera-Belmonte, A. Jochum, and A. Garcia, “Space-assisted irrigation management: Towards user-friendly products,” in Proc. ICID Workshop on Remote Sensing of ET for Large Regions, Montpellier, France, Sep. 17, 2003. [7] J. W. Hornbuckle, N. Car, E. W. Christen, T. Stein, and B. Williamson, “IrriSatSMS: Irrigation Water Management by Satellite and SMS—A Utilization Framework,” CSIRO, Griffith, Australia, Land and Water Science Report No. 04/09, 2009. [8] L. Johnson, R. Nemani, F. Melton, A. Michaelis, P. Votava, D. Wang, and T. Trout, “Information technology supports integration of satellite imagery with irrigation management in California’s central valley,” in Proc. 5th National Decennial Irrigation Conf., Phoenix, AZ, 2010, Paper No. IRR10–9645, ASABE, Publ. 711P0810cd. [9] N. J. Car, E. W. Christen, J. W. Hornbuckle, and G. A. Moore, “Using a mobile phone short messaging service (SMS) for irrigation scheduling in Australia—Farmers’ participation and utility evaluation,” Computers and Electronics in Agriculture, vol. 84, pp. 132–143, 2012. [10] B. Temesgen, S. Eching, B. Davidoff, and K. Frame, “Comparison of some reference evapotranspiration equations for California,” J. Irrigation and Drainage Eng., vol. 131, pp. 73–84, 2005. [11] Q. J. Hart, M. Brugnach, B. Temesgen, C. Rueda, S. L. Ustin, and K. Frame, “Daily reference evapotranspiration for California using satellite imagery and weather station measurement interpolation,” Civil Eng. Environ. Syst., vol. 26, pp. 19–33, 2009. [12] R. Allen, L. Pereira, D. Raes, and M. Smith, “Crop evapotranspiration: Guidlines for computing crop water requirements,” U.N. Food & Agriculture Org., Rome, Italy, FAO Irrigation and Drainage Paper #56, 1998. [13] D. D. Parker and D. Zilberman, “The use of information services: The case of CIMIS,” Agribusiness, vol. 12, pp. 209–218, 1996. [14] D. D. Parker, D. Zilberman, D. Cohen, and D. Osgood, “The economic costs and benefits associated with the California Irrigation Management Information System (CIMIS),” California Dept. Water Resources, Sacramento, CA, 1996. [15] USDA-NASS, Farm and Ranch Irrigation Survey, Table 36—Methods Used in Deciding When to Irrigate USDA National Agricultural Statistics Service, 2008. [16] K. Frame and B. Temesgen, “California irrigation management information system,” in Proc. Western States Water Council, Water Management Symp., San Diego, CA, Nov. 17–19, 2009. [17] W. G. M. Bastiaanssen, M. Menenti, R. A. Feddes, and A. A. M. Holtslag, “A remote sensing surface energy balance algorithm for land (SEBAL). 1. Formulation,” J. Hydrol., vol. 212–213, pp. 198–212, 1998. [18] R. Allen, M. Tasumi, and R. Trezza, “Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC),” J. Irrigation and Drainage Eng., vol. 133, pp. 380–394, 2007. [19] R. Allen, M. Tasumi, A. Morse, and R. Trezza, “A LandSat-based energy balance and evapotranspiration model in western US water rights regulation and planning,” Irrigation and Drainage Syst., vol. 19, pp. 251–268, 2005. [20] J. H. Kjaersgaard, R. G. Allen, M. Garcia, W. Kramber, and R. Trezza, “Automated selection of anchor pixels for Landsat based evapotranspiration estimation,” in Proc. World Environmental and Water Resources Congress 2009, ASCE Conf. Proc., 2009, doi:10.1061/41036(342)442. [21] R. Nemani, H. Hashimoto, P. Votava, F. Melton, W. Wang, A. Michaelis, L. Mutch, C. Milesi, S. Hiatt, and M. White, “Monitoring and forecasting ecosystem dynamics using the terrestrial observation and prediction system (TOPS),” Remote Sens. Environ., vol. 113, pp. 1497–1509, 2009.

1720


[22] R. L. Snyder, D. Spano, and K. T. Pawu, “Surface renewal analysis for sensible and latent heat flux density,” Boundary-Layer Meteorology, vol. 77, pp. 249–266, 1996. [23] R. Nemani, P. Votava, A. Michaelis, F. Melton, and C. Milesi, “Collaborative supercomputing for global change science,” Eos Trans. AGU, vol. 92, pp. 109–110, 2011. [24] M. C. Anderson, J. M. Norman, J. R. Mecikalski, R. D. Torn, W. P. Kustas, and J. B. Basara, “A multi scale remote sensing model for disaggregating regional fluxes to micrometeorological scales,” J. Hydrometeorol., vol. 5, pp. 343–363, 2004. [25] C. Neale, W. Bausch, and D. Heeremann, “Development of reflectancebased crop coefficients for corn,” Trans. ASAE, vol. 32, pp. 1891–1899, 1989. [26] C. Neale, H. Jayanthi, and J. L. Wright, “Crop and irrigation water management using high resolution airborne remote sensing,” in Proc. 54th IEC Meeting of the International Commission on Irrigation and Drainage (ICID)Workshop Remote Sensing of ET for Large Regions, Montpellier, France, Sep. 17, 2003. [27] D. J. Hunsaker, E. M. Barnes, T. R. Clarke, G. J. Fitzgerald, and P. Pinter, “Cotton irrigation scheduling using remotely sensed and FAO-56 basal crop coefficients,” Trans. ASAE, vol. 48, p. 13, 2005. [28] E. Rafn, B. Contor, and D. Ames, “Evaluation of a method for estimating irrigated crop-evapotranspiration coefficients from remotely sensed data in Idaho,” J. Irrigation and Drainage Eng., vol. 134, pp. 722–729, 2008. [29] P. Gowda, J. Chavez, P. Colaizzi, S. Evett, T. Howell, and J. Tolk, “ET mapping for agricultural water management: Present status and challenges,” Irrigation Science, vol. 26, pp. 223–237, 2008. [30] J. G. Masek, E. F. Vermote, N. E. Saleous, R. Wolfe, F. G. Hall, K. F. Huemmrich, G. Feng, J. Kutler, and L. Teng-Kui, “A Landsat surface reflectance dataset for North America, 1990–2000,” IEEE Geosci. Remote Sens. Lett., vol. 3, pp. 68–72, 2006. [31] E. F. Vermote, N. Z. El Saleous, and C. O. Justice, “Atmospheric correction of MODIS data in the visible to middle infrared: First results,” Remote Sens. Environ., vol. 83, pp. 97–111, 2002. [32] C. J. Tucker, “Red and photographic infrared linear combinations for monitoring vegetation,” Remote Sens. Environ., vol. 8, pp. 127–150, 1979. [33] F. Gao, J. Masek, M. Schwaller, and F. Hall, “On the blending of the Landsat and MODIS surface reflectance: Predicting daily Landsat surface reflectance,” IEEE Trans. Geosci. Remote Sens., vol. 44, pp. 2207–2218, 2006. [34] C. Boryan, Z. Yang, R. Mueller, and M. Craig, “Monitoring US agriculture: The US Department of Agriculture, National agricultural statistics service, cropland data layer program,” Geocarto International, vol. 26, no. 5, pp. 341–358, 2011. [35] L. F. Johnson and T. J. Trout, “Satellite NDVI assisted monitoring of vegetable crop evapotranspiration in California’s San Joaquin valley,” Remote Sens., vol. 4, pp. 439–455, 2012. [36] T. Trout, L. Johnson, and J. Gartung, “Remote sensing of canopy cover in horticultural crops,” HortScience, vol. 43, pp. 333–337, 2008. [37] G. Asrar, M. Fuchs, E. T. Kanemasu, and J. L. Hatfield, “Estimating absorbed photosynthetic radiation and leaf area index from spectral reflectance in wheat,” Agron. J., vol. 76, pp. 300–306, 1984. [38] S. N. Goward and K. F. Huemmrich, “Vegetation canopy PAR absorptance and the normalized difference vegetation index: An assessment using the SAIL model,” Remote Sens. Environ., vol. 39, pp. 119–140, 1992. [39] L. Johnson and T. Scholasch, “Remote sensing of shaded area in vineyards,” HortTechnology, vol. 15, pp. 859–863, 2005. [40] R. Allen and L. Pereira, “Estimating crop coefficients from fraction of ground cover and height,” Irrigation Science, vol. 28, pp. 17–34, 2009. [41] J. E. Ayars, R. S. Johnson, C. J. Phene, T. J. Trout, D. A. Clark, and R. M. Mead, “Water use by drip-irrigated late-season peaches,” Irrigation Science, vol. 22, pp. 187–194, 2003. [42] L. E. Williams and J. E. Ayars, “Grapevine water use and the crop coefficient are linear functions of the shaded area measured beneath the canopy,” Agricultural and Forest Meteorology, vol. 132, pp. 201–211, 2005. [43] L. Johnson and F. Melton, “Satellite irrigation management support with the terrestrial observation and prediction system,” in Proc. ASPRS 18th William T. Pecora Memorial Remote Sensing Symp., Washington, DC, Nov. 14–17, 2011. [44] D. Spano, R. L. Snyder, P. Duce, and K. T. Paw U, “Estimating sensible and latent heat flux densities from grapevine canopies using surface renewal,” Agricultural and Forest Meteorology, vol. 104, pp. 171–183, 2000. [45] G. N. Geller and F. Melton, “Looking forward: Applying an ecological model web to assess impacts of climate change,” Biodiversity, vol. 9, pp. 79–83, 2008.

[46] J. S. Michalakes, J. Chen, L. Dudhia, J. Hart, J. Klemp, J. Middlecoff, and W. Skamarock, “Development of a next generation regional weather research and forecast model,” in 9th ECMWF Workshop on the Use of High Performance Computing in Meteorology, Singapore, 2001, pp. 269–276. [47] R. Allen, I. Walter, R. Elliot, T. Howell, D. Itenfisu, M. Jensen, and R. L. Snyder, The ASCE Standardized Refernce Evapotranspiration Equation, Final Report (ASCE-EWRI), Task Committee on Standardization of Reference Evapotranspiration of the Environmental and Water Resources Institute, 2005. [48] A. Munevar and F. Chung, “Modeling California’s water resource systems with CALSIM,” in Proc. 29th Annual Water Resources Planning and Management Conf., ASCE Conf. Proc., 1999, doi:10.1061/ 40430(1999)95. [49] M. Rayej, R. L. Snyder, M. N. Orang, S. Geng, and S. Sarreshteh, “CALSIMETAW and WEAP models for water demand planning,” in Proc. ICID 21st Int. Congress on Irrigation and Drainag, ICID Transactions No.30-A, 2011. [50] R. L. Snyder, M. N. Orang, S. Geng, J. S. Maytac, and S. Sarreshteh, “Simulation of evapotranspiration of applied water,” California Dept. Water Resources, Sacramento, CA, California Water Plan Update 2005, 2005. [51] M. Dinguirard and P. N. Slater, “Calibration of space-multispectral imaging sensors: A review,” Remote Sens. Environ., vol. 68, pp. 194–205, 1999.

Forrest S. Melton is a Senior Research Scientist at California State University, Monterey Bay, CA, and is based in the Ecological Forecasting Lab at NASA Ames Research Center in Silicon Valley. He specializes in applications of remote sensing and eco-hydrologic models to address a range of natural resource management challenges.

Lee F. Johnson is a Senior Research Scientist with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA, and is based at NASA Ames Research Center in Silicon Valley. He specializes in vegetation canopy evaluation by remote sensing methods, and has analyzed multispectral and hyperspectral imagery from a wide variety of imaging systems with current emphasis on agricultural applications.

Christopher P. Lund is a Research Scientist with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA. He specializes in measurement of land surface fluxes of carbon dioxide and water vapor, SVAT modeling, and water budgets in agricultural systems.

Lars L. Pierce is a Research Scientist with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Andrew R. Michaelis is a Senior Software Engineer with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Samuel H. Hiatt is a Software Engineer with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Alberto Guzman is a Software Engineer with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Diganta D. Adhikari is a Research Associate with the Center for Irrigation Technology at California State University Fresno, Fresno, CA.


1721

Adam J. Purdy is a Research Assistant with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Bekele Temesgen is Chief of the California Irrigation Management Information System, California Department of Water Resources, Sacramento, CA.

Carolyn Rosevelt is a Research Assistant with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Kent Frame is a Program Manager II with the Water Use and Efficiency Branch, Division of Statewide Integrated Water Management, California Department of Water Resources, Sacramento, CA.

Petr Votava is a Senior Software Engineer with the Division of Science and Environmental Policy at California State University, Monterey Bay, CA.

Thomas J. Trout is Research Leader of the USDA–Agricultural Research Service Water Management Research Unit, Fort Collins, CO.

Edwin Sheffner is Deputy Division Chief of the Earth Science Division, NASA Ames Research Center, Moffett Field, CA.

Rama R. Nemani is a Senior Research Scientist with the Earth Science Division, NASA Ames Research Center, Moffett Field, CA.