APPENDIX C

LSA-NC Science /Implementation Task Group Session II Report and Recommendations

The Task Group met on February 19 , 1997 at the Holiday Inn Express in St. Paul, MN. The group focused on the progress since its last report and on those areas which may need enhancements for the second Enhanced Seasonal Observing Period during the cold season of 1997/98 (ESOP-98).

C.1 Background

The GCIP Implementation Plan (IGPO 1994a) identified the following features of the LSA-NC as important to GCIP:

• Role of winter snow accumulation and spring snowmelt in the annual water budget.
  
• Large natural inertia in the water runoff system due to lakes.
  
• Minimal orographic effects for precipitation.
  
• Soil freezing and thawing.

The results from the LSA-NC Detailed Design Workshop held in October 1995 (IGPO 1995a) were used by the first meeting of the LSA-NC Science/Implementation Task Group held in March 1996 (IGPO 1996a) to develop more specific recommendations for GCIP activities in the LSA-NC during 1997 and 1998. The Task Group recommendations primarily addressed those scientific issues that relate to snow and frozen ground processes and that can take advantage of the existing infrastructure and ongoing projects. These scientific issues were organized around three themes with specific activities recommended for each theme:

• Land Surface Model Physics
  
• Land Surface Modeling of SubGrid- Scale Heterogeneity Effects
  
• Monitoring of the Land-Surface State

These three themes along with the specific activities recommended for each theme and the status of these recommendations are described in later sections.

C.2 LSA-NC Implementation

A fifth study (PI: A. Super) will accelerate the development of a snow accumulation algorithm for use with NEXRAD radars. This funding will support real-time testing of an algorithm in Minneapolis and the adaptation of this algorithm for application to the "NIDS" 5 dBZ resolution product. This work is expected to complement other data-related studies being carried out in the LSA-NC by Steve and Tom Carroll to develop better algorithms for estimating snow on the ground and work by G. Peck and P. Groisman aimed at improving the estimates of snow amount derived from climate station data.

C.3 Land Surface Model Physics

R3.1--- During ESOP-97, the data sets for the variables identified in Table C-1 should be collected at one or more sites. Based on information presented at the first task group meeting, the Rosemount Experimental site and the Bondville, Illinois Climate Network site were identified as particularly suitable and with minor improvements in measurement capabilities could meet all of the data requirements given in Table C-1.

Status: J. Baker reported on the Status of the Rosemount Site measurements for ESOP-97. The Rosemount Experiment Station is located 24 km south of St. Paul. All measurements described below are made in a 17 ha (40 acre) field located on the south side of 160th Street, approximately 0.8 km east of the station office. The field is currently planted with maize, and there is a permanent mast installed in the center, so fetch exceeds 180 m in all directions. Precipitation gauges are installed within a Wyoming-type enclosure located 80 m north of the main mast. Data are transmitted via buried cable to a computer housed in a nearby building, and subsequently retrieved via telephone for archival at the St. Paul campus of the University of Minnesota. Routine meteorological data (averaged or summed on 30 minute intervals) that are collected with automated instrumentation include the following variables:

VARIABLE	INSTRUMENTATION
Air temperature (2m) Relative humidity (2m) Windspeed (0.4, 0.8, 1.4, and 2m) Solar Radiation (incoming and reflected) Longwave radiation (incoming and outgoing) Net radiation Precipitation quantity Precipitation quantity Snow depth	Vaisala HMP35 Vaisala HMP35 RM Young cup anemometers Kipp & Zonen pyranometers Eppley pyrgeometers REBS Q7.1 net radiometer Qualimetrics heated tipping bucket Belfort weighing gauge CSI Ultrasonic sensor

The following soils data (averaged or sampled on 30 min intervals) are collected:

Soil heat flux (2.5cm) Soil temperature (8 depths from 2.5 cm to 1 m) Water content (8 depths, from 2.5 cm to 1 m)

REBS heat flux plate & thermocouple thermistors time-domain reflectometry

The following data are collected manually (approximately weekly, more frequently as needed):

• Snow depth by visual observation of 37 snow sticks arrayed in a circle of 90 m radius.
  
• Snow density by gravimetric sampling
  
• Soil water content by neutron probe (20 cm depths to 160 cm)

Turbulent flux data are also collected. Sensible and latent heat flux measurements were made during spring 1997 snowmelt by eddy covariance, with a CSI 1-D sonic anemometer and Krypton hygrometer. For winter 1997-1998, continuous measurements of each will be made by conditional sampling and/or eddy covariance.

S. Hollinger reported on the status of the Bondville Site measurements for ESOP-97 including the new surface flux measurement site near to the Bondville site. A Campbell Scientific snow depth sensor was installed at the Illinois Climate Network station near Bondville, IL in early January. There were three significant snow events, and two snow melts after the installation of the sensor. These were the major snow events of the winter of 1996-1997 in the region. The first snow event began on 10 January resulting in a snow accumulation of approximately 5 cm. A second snow event occurred on 15 January resulting in a final total snow pack of 16 cm. The snowmelt occurred as a result of a warm front and warm air rain. All of the snow was melted in a period beginning at 1200 on 20 January through 1200 on 22 January when the snow pack decreased from 15 cm to 0 cm. On 26 January the third snow event began resulting in a snow accumulation of approximately 2.5 cm. This snow cover remained until 31 January when it melted as a result of a bright sunny day when air temperatures increased to 5C. At the flux station located east of Bondville, soil temperatures were colder in the no-till soybean residue than at the Bondville site which was under a heavy grass cover.

R3.2 ---The above datasets should also be collected during ESOP-98 at the same sites. Other suitable sites such as the Walnut Creek and Shingobee River watersheds can be considered for additional data collection efforts.

Status: Plans are proceeding to include both the Rosemount and Bondville sites for special data collection efforts during ESOP-98.

D. Rosenberry reported on the status of the measurements at the Shingobee River watershed and the site at Bemidji, MN. Soil-temperature and soil-moisture sensors were installed January 31, 1997 at the Bemidji Toxic Substances Hydrology site in northern Minnesota to complete a data-collection package that represents soil conditions in the center of a 2 ha field in a jack pine forest. Time-domain reflectometry (TDR) sensors and thermocouple thermometers already were installed at the Bemidji site in a related effort by the U.S. Geological Survey at depths of 50, 100, 150, 200, 250 and 300 cm below land surface. The January installation of identical probes at depths of 10, 20 and 30 cm below land surface meets the needs of GCIP by providing shallow soil-moisture data. This also will allow documentation of soil-moisture changes as the snowpack melts this spring. TDR probes and thermocouple thermometers also will be installed this June at the U.S. Geological Survey Interdisciplinary Research Initiative (IRI) site 65 km south of the Bemidji Site. These sensors will be installed in a mixed coniferous-deciduous forest at depths of 5, 15, 25, 35, 50, 100, 150, 200, 250 and 300 cm below land surface. In addition, a thermocouple will be installed at land surface at both the Bemidji and IRI sites. All sensors will be connected to Campbell Scientific dataloggers which will compute hourly and daily average values for soil moisture and soil temperature. TDR probes will be connected to a Tektronics cable testor, which when combined with the datalogger will process and convert TDR-probe signals to soil-moisture data. Calibrations will be made with a neutron soil-moisture probe inserted in neutron access tubes installed adjacent to the sensor strings, and with bulk-density measurements of nearby soil samples.

The measurements at the Walnut Creek watershed are made by the USDA/ Soil Tilth Laboratory. A NOAA wind profiler site is on the southern edge of this watershed. In addition, the Des Moines WSR-88D radar is nearby and could provide wind information. These data are part of the ESOP-97 data set. Further investigation is needed to determine whether all the critical measurements shown in Table C-1 will be available during ESOP-98.

R3.3 --- An evaluation of land surface models should be undertaken, possibly as part of a PILPS initiative, based on the data collected at all of the selected sites.

Status: No action has yet been taken on this recommendation.

Table C-1. Variables Required for Land/Hydrology Model Studies

Forcing measurements (30 min res) U component wind speed at 10 m V component wind speed at 10 m Temperature at 2 m Specific humidity at 2 m Surface pressure Surface skin temperature Precipitation quantity and type Surface Radiation - downward shortwave Surface Radiation - downward longwave

Validation Surface Radiation - upward longwave Surface Radiation - net radiation (measured) Streamflow Soil moisture (profiles) Soil temperature (profiles) Surface latent heat flux Surface sensible heat flux

Set up for Experiment Vegetation type and characteristics Site Site Description Surface Radiation - upward shortwave (albedo) Soil characteristics Wilting point Rooting zone Field capacity

C.4 Land Surface Modeling of SubGrid- Scale Heterogeneity Effects

Status: No action has yet been taken on this recommendation.

C.5 Monitoring of the Land-Surface State

The discussion and conclusions of the Taskgroup resulted in the follow-up recommendations given in Section C.6.

R5.2 --- Optimal methods to combine cooperative observer, satellite, and airborne gamma radiation snow data should be developed. These methods should produce snow fields with acceptable accuracy both for research studies (when all data can be used) and for operational applications (when only a subset of cooperative observer data are available).

Status: T. Carroll gave the special meeting on snowfall measurements adjustments a presentation and demonstration of the operational procedures and the developments in process to prepare maps of snowfall and snow water equivalent from an integrated set of data.

R5.3 --- The GCIP office should investigate whether more applicable radar algorithms like those to be used at the Minneapolis WSR-88D radar can be implemented operationally before ESOP-98 for those radar systems covering the LSA-NC.

Status: A report by A. Super and E. Holroyd III entitled "Snow Accumulation Algorithm for WSR-88D, Version 1" was sent to the participants at the Snow Measurement Adjustment meeting.. In addition, A. Super participated in the special meeting on snow measurement adjustments and presented a report on his work. It was agreed that there was little or no chance that a WSR-88D algorithm would be implemented operationally before ESOP-98 in any portions of the WSR-88D operational network.

R5.4 --- WSR-88D radar data from the Minneapolis site should be archived for ESOP- 97 and ESOP-98. Studies of snow water variability using these and other relevant data should be encouraged.

Status: The Level II data from each of the WSR-88D sites in the operational network are routinely archived at the NCDC in Asheville, NC. However, the cost to retrieve these data is such that GCIP could only afford to retrieve limited samples from this archive.

The GCIP/Data Collection and Management (DACOM) committee has taken the initiative to collect the data from the operational NEXRAD Information Dissemination Service (NIDS) for nine sites in the LSA-NC. However, there is some concern about the utility of the reflectivity data to derive estimates of snowfall.

A. Super agreed to use some NIDS data samples from the Minneapolis site to be supplied by S. Loehrer to derive estimates of snowfall for cases in which such estimates were previously derived from the full reflectivity data.

R3.5 --- The development of methods to combine remotely sensed and in situ soil moisture should be encouraged. Of particular interest are methods that are accurate at the beginning of the cold season, just before the soil freezes and snow cover commences, and just after snow cover has disappeared.

Status: T. Carroll showed results of soil moisture comparisons between aircraft and in-situ measurements made at the beginning of the cold season. It was agreed that GCIP should continue to encourage the development of methods to combine these two types of measurements.

R3.6--- To the extent possible within fiscal constraints, the GCIP should encourage and support routine soil moisture measurements at several sites within LSA-NC.

Status: J. Leese reported on some GCIP activities to support soil moisture measurements in the LSA-NC. Partial support was provided to the Water Resources Division of the USGS to install soil moisture sensors at the Shingobee River watershed. The surface flux site installed near Bondville, IL includes soil moisture sensors. J. Baker informed the meeting about plans to install soil moisture sensors at Lamberton and Waseca, MN.

J. Leese also reported on the GCIP activity to establish a North - South Transect of soil moisture and other measurements along or near 96W longitude. He particularly noted the contribution from the USDA/NRCS to make the data from their Soil Moisture/Soil Temperature Pilot Project available to GCIP. Noteworthy, for the N-S Soil Moisture Transect was the fact that the NRCS Project replaced the soil moisture sensors for three sites along the transect at a high priority in their schedule. The N-S transect will start at Plainview , TX (~30N latitude ) and continue North to Shingobee Watershed (~47N latitude) . Although sparse in the LSA-NC portion, the temporal variability of the soil moisture and soil temperature profiles over the course of an annual cycle should still be informative , especially during the cold period of the ESOP-98 from October 1997 to May 1998.

R 3.7 --- Satellite estimates of fractional snowcover should be obtained for the surface sites of interest (Rosemont, Bondville, Walnut Creek, Shingobee, etc.)

Status: Such estimates can be made from the operational meteorological satellite data acquired from either the NOAA polar-orbiting or the GOES satellites. However, the resolution is likely not sufficient to derive estimates of fractional snowcover for specific sites.

J. Leese reported on a Winter Cloud Experiment (WINCE) conducted by W. Smith at the University of Wisconsin. WINCE made use of the NASA ER-2 aircraft to fly the MODIS Airborne Simulator (MAS) for calibration checks. Arrangements were made with W. Smith to schedule some of the flights between Madison, Wisconsin and Bondville, IL during a two-week period in February, 1997. MAS data from the flights over Bondville are expected to become available in the near future. The MAS is an airborne sensor which has most of the remote sensing characteristics of the Moderate resolution Imaging Spectroradiometer , or MODIS to be flown aboard the EOS AM-1 satellite scheduled for launch in 1998. The MODIS data could be important for GCIP in the later stages of the five-year Enhanced Observing Period.

C.6 Recommendations

The following recommendations primarily address collection and monitoring efforts to enhance ESOP-98 data sets, and are identified as Recommendation 98-x.

C.6.1 Data Collection to Support Land-Surface Model Development.

Enhanced data collection efforts are planned for the Rosemount and Bondville sites. This will provide a core set of data for model development. However, there are opportunities for enhancement. In particular, there are ongoing data collection efforts in the Walnut Creek watershed that may meet many, if not all, the requirements shown earlier in Table C-1. It may be possible to meet all requirements with little effort and funds. Its location, roughly intermediate in latitude between Rosemount and Bondville, would provide a worthwhile enhancement to the GCIP data base.

Recommendation 98-1: The GCIP office should investigate whether all of the critical measurements in Table C-1 will be available from the Walnut Creek watershed. If only minor enhancements are needed to meet all requirements, we recommend that the GCIP office explore options to achieve those enhancements.

C.6.2 Data Collection Efforts for Modeling of Subgrid-Scale Heterogeneity.

Enhanced data collection efforts to document subgrid-scale heterogeneity are planned for Bondville, an area of generally ephemeral cold season snow cover. It would be desirable to also have suitable data collection efforts in the northern portions of the basin with longer lasting snow cover. Such data collection efforts would need to be on the scale of numerical weather prediction (NWP) grid and hydrologic model scales. We recognize that this can be costly and difficult and may not be possible within fiscal constraints. Nevertheless, its importance in accurate modeling of the land-surface state is sufficiently great that we make the following recommendation.

Recommendation 98-2: The GCIP office should investigate options for the collection of data on the subgrid-scale heterogeneity of snow cover at one or more sites in the northern portions of the basin. It may be most economical to utilize remotely sensed observations from satellites and aircraft. However, for purposes of model development, it would be advantageous to incorporate some high spatial resolution ground-based measurements of snow cover and water equivalent.

C.6.3. Monitoring of Snow Water Equivalent

The report by Eugene Peck on the snow measurements program in the LSA-NC included a number of recommendations. The Taskgroup generally agrees with these recommendations. Thus, the recommendations of the Taskgroup, largely adapted from the Peck report, are as follows:

Recommendation 98-3.1: The GCIP program office should take actions to improve snow records for ESOP-97 and ESOP-98. In order of priority, these actions are as follows:

98-3.1.1. A comparison of the Groisman and Peck factors to characterize site exposure should be undertaken. It would be advantageous if the two approaches provided comparable results because the Groisman approach only requires station history information and can be applied to many more stations than the Peck approach which requires a detailed knowledge of the site. However, the Peck approach is presumably the more accurate.

98-3.1.2. An application of any method for making exposure adjustments to snow records requires a reasonably accurate estimate of the wind. This presents a serious problem for climatological stations not near a synoptic station and may also be a problem when a synoptic station's data quality is suspect, an example of which is reported in the Peck report. The NOAA ETA model provides wind movement at a 10 meter height at a grid resolution of 40 km. This internally consistent wind data set at an adequate spatial resolution provides a potential solution. The Taskgroup recommends that GCIP support or arrange for a test study that would use ETA model winds to adjust snow records at climate stations and to evaluate the reliability of wind records at synoptic stations in the LSA-NC region.

98-3.1.3. The GCIP office should arrange that observed and adjusted values of snowfall for selected climate stations in the LSA-NC be released in near real-time.

98-3.1.4. The GCIP office should encourage activities to improve mapping of snow depth, coverage, and water equivalent, using a combination of airborne gamma and in situ measurements.

Recommendation 98-3.2. The GCIP project office should foster actions to improve historical snow records in the following priority order:

98-3.2.1. The GCIP office should encourage the development of isohyetal winter and annual maps of snowfall in map and gridded data format for the northern portions of the LSA-NC. Current maps are inadequate.

98-3.2.2. In order to provide for a consistent record and avoid discontinuities at the international boundary, it is recommended that the Canadian snow records be adjusted on a daily basis using similar techniques. Both observed and adjusted values should be published.

Recommendation 98-3.3. The problems associated with snow records in the LSA-NC also affect records for other areas of the U.S. The GCIP office is encouraged to work with the NWS, NCDC, and other relevant organizations to improve all U.S. snowfall records. Specifically, snowfall records should be adjusted in near-real-time on an operational basis. Both observed and adjusted values should be published.

C.6.4 Monitoring of Soil Moisture

Recommendation 98-4.1. The GCIP office should continue to encourage the development of methods to combine remotely sensed and in situ soil moisture measurements.

Recommendation 98- 4.2. The Taskgroup supports the GCIP office's initiative to establish a north-south transect of soil moisture monitoring sites and encourages its continued development.

C.6.5 Monitoring of Surface Albedo

Despite the importance of albedo in determining the coupling between the land surface and the atmosphere and the large albedo changes associated with snow accumulation and melt, there are no operational plans to measure this variable on an area-averaged basis, because of the high costs associated with airborne platforms. However, this is an important component to address.

Recommendation 98-5. The GCIP office should investigate opportunities to obtain large area observations of albedo, such as was done during the ESOP-97 with ER-2 flights of the MODIS Airborne Simulator.