User Tools

Site Tools



This shows you the differences between two versions of the page.

Link to this comparison view

west_stationdata:analysislog_2 [2012/06/14 02:33]
— (current)
Line 1: Line 1:
-====== Soil moisture analysis log ====== 
-This log mainly covers analysis of SNOTEL soil moisture data. 
-**See also:** The [[west_stationdata:​analysislog_1|soil temperature analysis log]], and the [[west_stationdata:​overview|project overview]]. 
-**Feb 21, 2012** - A couple ideas: 
-  * Mean soil moisture the week before snowpack onset and the day of snowpack onset. 
-    * Does this correlate with winter soil moisture? Until when (relative to peak snowpack or snowmelt)? 
-  * Does day of snowmelt correlate with late summer soil moisture? 
-    * In monsoon vs non-monsoon areas? Perhaps monsoon areas are more correlated with summer precip? 
-**Feb 16, 2012** - Have done lots in the last few days: 
-  * **New 30-year mean data** (SNOTEL SWE, SNOTEL Precip, and SnowCourse SWE) downloaded from NRCS website, error checked, and tabulated in a new spreadsheet. See ''​rawdata/​longterm_averages/​SnowSurvey_7100Avg_Master.gnumeric''​. 
-    * Makes some older spreadsheets obsolete, though there are some old copies. 
-    * SWE and precip data are exported as .csv files to the curated data folder for use in data-analysis scripts 
-  * **New station inventory spreadsheet** is located at ''​rawdata/​station_inventory/​NRCS_inventory_MASTER.gnumeric''​. 
-    * Created by downloading and editing .csv files from the NRCS [[http://​​nwcc/​inventory|data search]] webpage. 
-    * This file is being used to keep track of what SNOTEL/​Scan/​SnowCourse data is available for snowpack and soil profiles, and what data has been downloaded already. 
-    * Might replace the other inventory .csv files in this directory (some scripts still rely on these though). 
-  * Both of these files have some documentation available in the first sheet. 
-  * Some general reorganizing and cleaning out of old data and scripts in the rawdata, and data_analysis folders that contain this project. 
-    * ''​obsolete_scripts''​ holds old SNOTEL network analysis scripts that may still be useful but are being replaced by new ones. There is a corresponding collection of data files in the rawdata directory. 
-    * ''​curated_data''​ holds textfiles that are output from spreadsheets or scripts that operate on raw data (as it might be useful to version these). 
-  * New script to calculate many metrics of climate and interannual variability (in snow, precip, temp) from the daily data. It outputs a large dataset to the ''​curated_data''​ directory and is read by plotting scripts. **Will be documented [[programdocs|here]]** 
-  * Downloaded a large amount of new data from the NRCS website to ''​allsensors_daily'',​ and ''​soilsensors_hourly''​ directories. 
-    * Now have daily data for all SNOTEL sites with soil profiles from install to 2010 (working on 2011) 
-    * AZ, NM, UT, and CO have complete sets of the SNOTEL soil profile data (NV, WY, ID, MT to go still). 
-  * Met with Dave today, and we agreed that **more daily data is needed** - preferably all of it for the last 10 years or so. 
-**Feb 3, 2012** 
-{{ :​west_stationdata:​plot_4sites_distrib.png?​300|Sites in each column are Trial Lake, BL Trail, Chepeta, Little Grassy (Right to left)}} 
-  * plot_monthly_sm_distrib.m has been reworked and renamed to plot_soilsensor_distrib.m. It is a bit more generic and will plot soil temperature data in the same fashion. 
-  * Have been working on some case-study type plots comparing high/low elevation/​swe sites (in a factorial way). Results for Trial Lake, Ben Lomond Trail, Chepeta, and Little Grassy are **to the right** ->. Chepeta seems to defy expectations a bit by being highly variable in spring. (Plot generated by plot_4sites_distrib.m) 
-  * Met with Dave today and we agreed that a main message of this analysis so far is that soil moisture is highly variable inter-annually,​ between sites. and even within seasons. There are a few somewhat predictable patterns though: ​ 
-    * Low elevation sites seem to show more summer drought 
-    * Soil moisture stays constant below winter snowpacks, but there is large inter-annual variability between winters. This probably depends on year-to-year differences in antecedent (fall) precip. 
-    * In general, soil moisture variability is higher at low elevation sites (especially outside summer months). This could be shown with a CV vs elevation plot perhaps. 
-  * **TO DO:** 
-    * Explore the soil moisture CV vs elevation relationship (see above) 
-    * Need greater power for analysing case studies, including binning sites into climate-space envelopes. First this requires that we define ranges in elevation, MAT, MAP, SWE, melt timing, summer precip, etc, across the SNOTEL network. Second, how could we group sites according to particular criteria (High elevation sites with low snowpacks for instance.) 
-      * Such data could be used to select appropriate sites for case studies (such as replacing Chepeta in the figure at right). 
-    * Some of the temperature data from PRISM should probably be used in climate-space analysis rather than SNOTEL temp data. 
-    * There may be some interesting soil moisture patterns at monsoon sites that should be explored. The shape of the precip accumulation line might be useful for defining sites w/ summer rain. 
-    * Collect data from more sites - MT, AZ, NM are left. 
-**Jan 28, 2012** 
-  * There are now some new testing functions in the SNOTELdata project. One that plots two datasets to view changes (useful for comparing filtered/​unfiltered,​ or transformed/​untransformed datasets) and one that will generate SNOTEL datsets with original values, or with missing columns or dummy values. 
-  * The plot_monthly_sm_distrib.m file has been updated to show an all-years plot. 
-  * Dave and I discussed doing some side-by-side comparisons of the quarterly histograms from several sites that would illustrate low/high elevation and snowpack scenarios. 
-**Jan 21-26, 2012** 
-{{ :​west_stationdata:​plot_monthly_sm_distrib-828wy2010.png?​300|plot_monthly_sm_distrib.m output for Trial Lake, WY2010}} 
-  * Created plot_monthly_sm_distrib.m to print a site's monthly histograms of all soil moisture data for each year in the analysis. 
-  * Have been working on some testing functions for loadsnotel.m and the data cleaning/​filtering routines. 
-  * Possible test cases that involve replacement/​addition of sensors between years are site 396 (Ts -2 disappears after 2004), and 435 (Ts -20 disappears after 2004). 
-**Jan 18, 2012** - Some initial goals: 
-  * Each sensor needs to be normalized to 0-1 (after data cleaning) 
-  * Plot histograms of the data during different time periods(months) of the year (some will be bi-modal) 
-  * size of histogram modes indicate time sensor spends at particular value 
-  * try plotting in 3d - see figure 6 in bowling_11 paper 
-  * Next week - soil moisture distributions for 1 site, all months, 1 sensor depth. 
west_stationdata/analysislog_2.1339655584.txt.gz ยท Last modified: 2012/06/14 02:33 by gremau