2016 Land Use Data for Model

nickrsan commented 7 years ago

@josue-medellin @andybell This thread is for reproducing the land use data for DWR. After spending a while wondering just what units DWR was using, I realized how to reproduce their land use units. Using the modified 2015 data Quinn published, I'm able to reproduce his values per DAU/County to within a reasonable margin (eg: Morrie's copy has 6779 acres of alfalfa, and the data I generated has 6787 acres) for all land use types except Riparian and Wet Herbaceous. However, the total of those two in the copy sent by Morrie and the data I generated comes out very close. I suspect that maybe the data Morrie had was generated with older land use info, but haven't investigated yet.

Currently, I'm planning to produce the 2016 data with this method, but won't send it unless you two agree this is sufficient. I can also investigate the sources of error before sending, but suspect most of the variation except the Riparian/Wet Herbaceous issue is due to methodology differences between Quinn's summary and min.

josue-medellin commented 7 years ago

I think it works fro our needs Nick unless @andybell thin is not sufficient. Wetland herbaceous wont go in our calculations.

Kind regards,

Josue

On Tue, May 16, 2017 at 4:56 PM, Nick notifications@github.com wrote:

@josue-medellin https://github.com/josue-medellin @andybell https://github.com/andybell This thread is for reproducing the land use data for DWR. After spending a while wondering just what units DWR was using, I realized how to reproduce their land use units. Using the modified 2015 data Quinn published, I'm able to reproduce his values per DAU/County to within a reasonable margin (eg: Morrie's copy has 6779 acres of alfalfa, and the data I generated has 6787 acres) for all land use types except Riparian and Wet Herbaceous. However, the total of those two in the copy sent by Morrie and the data I generated comes out very close. I suspect that maybe the data Morrie had was generated with older land use info, but haven't investigated yet.

Currently, I'm planning to produce the 2016 data with this method, but won't send it unless you two agree this is sufficient. I can also investigate the sources of error before sending, but suspect most of the variation except the Riparian/Wet Herbaceous issue is due to methodology differences between Quinn's summary and min.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ssj-delta-cu/ssj-calsimetaw/issues/6, or mute the thread https://github.com/notifications/unsubscribe-auth/ANGlbEwe3NSkpWhQTiKG2k1pUq0jT5c1ks5r6jebgaJpZM4NdNw6 .

-- Josué Medellín-Azuara, Ph.D. Associate Research Engineer Department of Civil and Environmental Engineering Center for Watershed Sciences (map https://goo.gl/maps/uce4oivhedt) University of California, Davis One Shields Ave. Davis, CA 95616 Mobile: + 1 530 574 8019 Website: http://watershed.ucdavis.edu/medellin

Past-Convener California Water and Environmental Modeling Forum, CWEMF http://cwemf.org

Be green! Please don't print this e-mail unless you really need to.

nickrsan commented 7 years ago

OK, here's the summary table, Andy, once we discuss tomorrow, I'll send it to Morrie. 2016_land_use_summary.xlsx

andybell commented 7 years ago

Looks good to me. Not sure when the initial area totals were calculated but it seems plausible that the totals were from the initial release of the LandIQ 2015 data which would result in slight discrepancies.

nickrsan commented 7 years ago

OK, new info:

I'm attaching a new version of the spreadsheet - the data hasn't changed, but I've included a new sheet that shows a comparison of the values for the 2015 data between the method used last year and the method I used this year on the 2016 data. I included percent errors, which are quite small, with two caveats:

Land uses with very few acres have higher errors, which makes sense - more subject to edge effects in the analysis
- Wet Herbaceous and Riparian are very far off, but when you sum them together and compare (see last line in sheet), they have very low errors. I'm using Quinn's latest land use file from last year, and I suspect that it had different assignments for those categories than what was used in the model, but I haven't verified by checking an earlier version.

2016_land_use_summary.xlsx

As for the methodology, I used Polygon to Raster in ArcGIS to convert the 2015 land use data to raster, using the 2016 land use data Andy produced as the snap raster. I then used Tabulate Area to get the values, exported the table to Excel and added the crop categories and Pivot Table.

andybell commented 7 years ago

@nickrsan I ran the following query in postgis to get the level 2 area in sq meters per crop.

create or replace view level_2_ct_wy2016 as 
select * from crosstab(
    'select level_2,dauco,sum(sum)::integer from calsimetaw.dau_landuse_wy2016 group by 1,2 order by 1,2',
    'select distinct dauco from calsimetaw.dau_landuse_wy2016 order by 1'
) as ( 
    level_2 text,
    "DAUCo18501GA" integer,
    "DAUCo18507GA" integer,
    "DAUCo18539GA" integer,
    "DAUCo18634GA" integer,
    "DAUCo18648GA" integer,
    "DAUCo18657GA" integer
);

I'm getting slightly different numbers than the values calculated in ArcGIS. There are a few new crop categories in 2016 that are missing from the spreadsheet (Asparagus, Carrots, Eucalyptus, etc).

level_2_ct_2016_comparison_TabArea.xlsx

ssj-delta-cu / ssj-calsimetaw

2016 Land Use Data for Model #6