APPENDIX E: DATA PROCESSING AND STORAGE 



Eve Porter* University of Alaska 



Bob Porter and staff Data Processing Section, Geophysical Insti- 

 tute, University of Alaska 



In order to process the immense amount of information collected in a 

 program such as this, a considerable effort went into data handling. 

 Data processing at the University of Alaska was initiated in June 1970. 

 Staff of the Geophysical Institute's data processing section were re- 

 sponsible for this phase of the program. The objectives were to make 

 processed data available to the investigator shortly after acquisition 

 of the raw input and to provide statistical and other analytical manipula- 

 tions of the data so that early inferences could be drawn from the studies 

 and comparisons made among subprojects. Unfortunately the design of the 

 data system began concurrently with the field program without the lead 

 time necessary to design the optimal system. This was necessitated by 

 lack of sufficient funds early enough in the 1970 program. As a result, 

 the first goal of rapid turn-around was achieved by the middle of the 

 summer for most Barrow subprojects. The second goal is still under pur- 

 suit. Limited preparation time, personnel, and funds required setting 

 a priority on data processing. Only the Barrow projects which shared 

 common sample periods and sites received top priority treatment in the 

 data processing system. Other projects are now being serviced. By next 

 summer all projects should have equal access to the data processing 

 services . 



The first step in the process was the design of format input sheets 

 for each subproject. Field data were transferred to these format sheets 

 at Barrow and mailed to College. The data were edited and key punched. 

 As programs became available the data were processed. For optimal opera- 

 tions, a four-day turn-around was common during each 10-day sample interval, 



There are approximately 65 subroutines and 15 mainline processing pro- 

 grams, consisting of about 16,000 source cards. All programs are written 

 in FORTRAN IV or Assembler Language for the IBM 360 Model 40, and have 

 been compiled on DOS Release 21. All programs can be supplied as cards, 

 listings, card image tape, or as a restore tape for an IBM 360. The 

 system is designed with a view to later implementation as a random access 

 retrieval file in order to facilitate cross-correlations between the many 

 different kinds of data collected in the Tundra Biome . The data are 

 entered into the system on punched cards, each sample or sample group 

 being identified by a unique identification number, and each type of data 

 being assigned a "data type" number which is intended to meet U.S. IBP 

 interbiome data banks criteria. 



All data are heavily edited before being entered into the system for 

 processing, to assure where possible, that the date, site identification, 

 and the data itself are reasonably within the limitations of computer 



^''Principal author. 



139 



