WMOD00570

• Title: Precipitation Enhancement Effects of Sorghum, Using a Daily Growth Model
• Date: 4/1/1982
• Weather Modification - Doc Type: Report

The rank correlations between mean normalized precipitation and mean normal-
ized yield for the 11 scenarios were calculated for Lubbock and Goodland.
The coefficient for Lubbock is 0.87 and is 0.98 for Goodland, again showing
the strong relationship between additional precipitation and increased yield.
Most of the yi e 1 ds at the other three sites were in the noi se, and corre 1 a-
tion coefficients would have questionable significance.

Interpretation of the practical significance of these scenarios requires
assumptions about the noise level of the model (arbitrarily set at 5 percent)
and the yield increase that must be realized in an actual sorghum field
before it can be detected above the natural variations in agricultural
production. This latter value has also been set, arbitrarily, at 5 percent.
We can now estimate (1) the frequency with which each scenario can provide a
"real" increase in yield (i .e., 6 percent or more) and (2) the amount of
additional precipitation needed to achieve this "rea'" yield increase.

Only Scenarios 7A and 7B produced yield increases averaging 6 percent or more
at all five sites, and Scenario 6 produced "significant" average increases at
Goodland, Oklahoma City, and Lubbock. However, examination of the detailed
results show that, although a given scenario may have a small effect on yield
in the mean, it may still produce significant yield increases in some years.
Table 5 shows the number and percent of years that yield increases were above
the assumed noise level of 5 percent.

:.1.:

The above analyses give the average yield increase for each scenario. We
al so looked at the average preci pitat ion increase associ ated with a :6-percent
increase in yield. "To estimate this threshold precipitation amount, the
precipitation increases for all scenarios in years with 5-, 6-, and 7-percent
increases in yield were averaged for each site (table 6). These numbers were
compared with the mean scenario increases from table 4. For the purpose of
this exercise, it is assumed that a scenario is "effective" if the mean pre-
cipitation increase is equal to or greater than the mean precipitation needed
to achieve a 6-percent (average) yield increase. Scenarios 7A and 7B are
effective at all five sites, and scenarios 4B, 5B, 6, and 8 are effective at
four of the five sites. Scenarios lB, 2, and 5A are effective only at
Lubbock, and Scenarios 3A and 3B are ineffective at all sites.

Timing

Precipitation in particular periods during plant growth and development is
especially important. For grain sorghum, moisture stress is critical in the
period of floral bud development and very critical in the heading period
(Neild, 1975). These periods are within stages 5 and 6 of the sorghum model,
and table 3 shows the rank correlation coefficients between stage precipita-
tion and yield for all five sites. For stages 5, 6, and 5 and 6 combined,
the coefficients range from low to moderate; only 8 of the 15 values in the
table are statistically significant at the 5-percent level. These results
indicate that the crucial periods for additional precipitation are shorter
than, or perhaps overlap, the phenological stages of plant development.

To narrow the time span, we arbitrarily selected the 2-week period from
7 days before the start of grain filling to 7 days after the start. Lubbock
was used because it generally shows the greatest correspondence between

"

€-

18