# apparently inconsistent results depending on allocation selection for product system calculation.

194 views

I have several issues, but will focus on a single problem that I can't understand. I have a process with two products (pigs and manure, with value, hence receiving an allocated share of burdens). I noticed some inconsistencies (I could not reproduce the allocated result from the unallocated result), and made the following test:

1) create product system with process allocation set to physical and 100% burden applied to the pig output (5500 head, which is the reference flow in the process definition, and set to the quantitative reference product). Calculate the result (A, below).

2) edit the process with causal allocation where the manure nutrients co-product are assigned 100% of all the inputs and emissions. Change the quantitative reference to the manure nutrients. Create another product system and calculate for the reference flow of manure (here defined as 7007208 gallons). Result B, below.

I expected the two systems to return identical results (that is the GWP for 5500 head and for 7007208 gallons should have the same numerical result since the only change was to make the quantitative reference flow in each case receive all the burdens - ie no allocation). However, this is not the result. I performed this test because I could not reproduce the allocated (from product system with economic allocation defined in the process) result from the unallocated result for 5500 head.

A) Contribution tree GWP 5500 head, 100% allocation to hogs:

Contribution Process                                     Amount               Unit

100.00% market pig; USFRA BL (attr) - US-IA 2047560.05775907 kg CO2 eq

21.19% grow rations-BL (attr) - US-IA 433872.11790973 kg CO2 eq

17.00% piglets; deep pit stalls barn; at farm (attr) - US-IA 348071.86456576 kg CO2 eq

03.16% market for propane, burned in building machine |  | Cutoff, U - GLO 64634.80812051 kg CO2 eq

03.10% market group for electricity, medium voltage |   | Cutoff, U - US 63391.90269583 kg CO2 eq

03.01% market for transport, tractor and trailer, agricultural | l | Cutoff, U - GLO 61609.65595061 kg CO2 eq

01.76% market for liquid manure spreading, by vacuum tanker |   | Cutoff, U - GLO 36050.86202478 kg CO2 eq

01.68% swine; grow barn; slatted floor 34430.55116470 kg CO2 eq

00.21% mortality management; USFRA; disposal at farm - US 4216.84042364 kg CO2 eq

00.13% market group for tap water | tap water | Cutoff, U - GLO 2592.07215338 kg CO2 eq

00.10% market for diesel, burned in agricultural machinery |  | Cutoff, U - GLO 2042.77299611 kg CO2 eq

B) Contribution tree GWP 7007208 gal manure, 100% allocation to manure:

Contribution  Process                                 Amount                           Unit

100.00% market pig; USFRA BL (attr) - US-IA 2007281.62914145 kg CO2 eq **NB: process name, not product is shown here**

21.19% grow rations-BL (attr) - US-IA   425337.23635443 kg CO2 eq

14.99% piglets; deep pit stalls barn; at farm (attr) - US-IA      300946.36293650 kg CO2 eq

03.16% market for propane, burned in building machine |  | Cutoff, U - GLO 63363.34924720 kg CO2 eq

03.10% market group for electricity, medium voltage |   | Cutoff, U - US 62144.89339770 kg CO2 eq

03.01% market for transport, tractor and trailer, agricultural |   | Cutoff, U - GLO 60397.70540977 kg CO2 eq

01.76% market for liquid manure spreading, by vacuum tanker |  | Cutoff, U - GLO 35341.68965994 kg CO2 eq

01.68% swine; grow barn; slatted floor      33753.25330371 kg CO2 eq

00.21% mortality management; USFRA; disposal at farm - US    4133.88917670 kg CO2 eq

00.13% market group for tap water | tap water | Cutoff, U - GLO     2541.08236117 kg CO2 eq

00.10% market for diesel, burned in agricultural machinery |   | Cutoff, U - GLO 2002.58871311 kg CO2 eq

When I look at the inventory (tot. req) it is also different. (can't post due to size limit). What can cause the inventory to change when all the inputs and outputs are identical but only the quantitative reference is changed (but the full process output calculated for both situations - the results should be identical by all that I understand in LCA calculations).