3.1.3. Grower Perceptions and Attitudes

The data collected on farmer perceptions and attitudes towards groundwater availability issues are summarized in Table 4. Less than one-third of growers observed a change in their wells' depth to water while over two-thirds of respondents indicated they perceived there was not a problem with the groundwater supply at their farm or in the state. A test of independence in these responses indicated that they had a statistically significant dependence: those who perceived there had been a change in their wells depth to water were more likely to believe there was a groundwater problem in their farm or at the state level.

The U.S. Geological Survey (USGS) has recently published a map of the Potentiometric Surface of the Mississippi River Valley Alluvial Aquifer for the Spring 2016 [25] that shows the location and gradient of the aquifer's cone of depression. A cross-tabulation of farmer perceptions and their location in the center of the cone of depression is presented in Table 5. A Pearson test of independence of the responses showed evidence that farmers located in the cone of depression were more likely to observe a change in their well levels and think there was a groundwater problem at the farm or state level. Half of those located in the center of the cone of depression believed there was a groundwater problem as opposed to 29 percent amongst those located outside that area. Similarly, forty-six percent of those

in the cone of depression area observed a change in the depth-to-water in their wells while only 26 percent of those outside the area noticed such a change.


**Table 4.** Summary statistics of farmer perceptions and attitudes.

Note: GW is Groundwater.

**Table 5.** Aquifer "cone of depression" and farmer perceptions and attitudes.


#### *3.2. Probit Regression Analysis*

The estimated models of practice adoption fit the data relatively well with pseudo-*R*<sup>2</sup> (McFadden's) ranging between 0.156 to 0.545. Except for micro-irrigation (only five adopters), the conservation practices being analyzed had at least one factor with a statistically significant coefficient. The probit regression coefficients are detailed in Table 6 for all factors except cracking soils and income, which are detailed in Table 7.

Tailwater Recovery System (TWS) adoption was positively and significantly influenced by the farmers perception that a groundwater problem existed (GW prob.). The marginal effect (*dy*/*dx*) indicated that a producer who becomes aware of the groundwater problems in the Delta area would be associated with a 25 percent higher likelihood of adopting TWS. The data indicated that farmers who do not use groundwater for irrigation have not adopted TWS.

For OFWS, the number of irrigated hectares under operation (Irr.area) was positively and significantly associated with the adoption of OFWS. The calculated marginal effect indicated that for an additional 40 hectares of farmed land, the probability of a farmer adopting OFWS was 0.8 percent higher. The data indicated that farmers who did not use groundwater for irrigation have not adopted OFWS.


**Table 6.** Results from probit regressions (coefficients by income level in Table 7). Irr, Irrigation.

Note: Educ. is years of formal education; GW Prob. is an indicator variable for perception of a groundwater problem at the farm or state level; P. Cost is the cost of pumping; and Cons. pr. is participation in a conservation program. Standard errors are in parentheses. \*,\*\*,\*\*\*: significant at *p* < 0.1, *p* < 0.05, *p* < 0.01, respectively. (a) negative cases predict failure; (b) positive cases predict failure.

Computerized Hole Selection (CHS) was positively and significantly associated with the number of irrigated hectares, the perception of the existence of a groundwater problem, and having an income between \$100,000 and \$150,000. The marginal effects indicated a 0.5 percent higher probability of CHS adoption for an additional 40 ha of land irrigated, and the probability of adoption increased by 30 percent when a farmer realized there was a groundwater problem at the farm or state level.


**Table 7.** Results from probit regressions (continued) on "cracking" soils and income levels.

Note: base income is \$50,000 or less; k represents thousands of dollars. Standard errors are in parentheses. \*,\*\*: significant at *p* < 0.1, *p* < 0.05, respectively. (a) negative cases predict failure; (b) positive cases predict failure; (c) positive cases predict success.

The adoption of surge irrigation (surge) was positively and significantly associated with the perception of the existence of a groundwater problem and negatively by the pumping cost. The negative influence of the pumping cost variable was a departure from the hypothesized relations in the GAO report. Because this variable is a combination of various data with a fundamental rooting in the county of residence claimed by the farmer, there may be confounding of factors, the identification of which is beyond the scope of this study. However, the result was driven in part by the fact that nobody claiming to reside in the cone of depression (highest pumping cost) used surge irrigation. Surge irrigation is harder to manage in the cracking clays that are a common soil type in that area , but this effect did not appear statistically significant in this regression. The marginal effect calculations suggested that the probability of adoption of surge increased by 0.2 percent for an additional 40 hectares of irrigated land added to the operation, but an increase of one percent in the cost of pumping would decrease the adoption probability by 0.55 percent. The data indicated that farmers who do not use groundwater for irrigation have not adopted surge.

The use of SMS was significantly associated with irrigated area (positive at five percent), rice production (negative at five percent), participation in a conservation program (positive at five percent), and increasing income (from the baseline to the \$100 k to \$150 k income bracket, positive at five percent). The estimates suggested that an increase of 40 hectares in irrigated land would result in a two percent higher probability of adopting SMS; the choice of growing rice would reduce that probability by 50 percent, and the participation in a conservation program would add 45 percent to the SMS adoption

probability. The data indicated that farmers who do not use groundwater for irrigation have not adopted SMS.

With respect to micro-irrigation, the probit regressions did not find statistically significant effects. This may be due in part to the relatively few respondents who claimed to practice it. For center pivot irrigation, however, there was enough variability to show a statistically significant and positive effect of the number of years of farmer formal education. For every additional year of formal education completed, the farmer was 0.68 percent more likely to adopt center pivot irrigation. The data indicated that farmers who do not use groundwater for irrigation have not adopted center pivot.

The adoption of a pump timer was positively and significantly associated with the number of irrigated hectares and the perception that a groundwater problem existed at the farm or state level. An additional 40 hectares or irrigated land was associated with a 0.6 percent higher probability of adoption, and the realization that a problem with groundwater stock existed in the state implied a 41 percent higher probability of employing a pump timer. All farmers in the sample with incomes above \$250,000 used pump timers.

Flow meter adoption was also positively and significantly associated with the number of irrigated hectares. An additional 40 hectares or irrigated land were associated with a one percent higher probability of adoption. All farmers in the sample with incomes above \$200,000 used flow meters.

Growing cover crops was negatively and significantly associated with farmer experience as represented by the number of years farming and positively associated with the \$200,000 to \$250,000 income bracket. An additional year of farming experience was associated with a 1.1 percent lower probability of growing cover crops. Since the agronomic benefits of cover crops payoff over a longer time-horizon, a farmer getting closer to retirement may be less eager to invest in a practice for which she/he will not see most of the benefits.

In terms of identifying factors that are associated with the adoption of conservation practices, the results indicated the following: irrigated area (positive), GW use (positive), rice (negative), years farming (negative), education years (positive), perception of GW problem (positive), pumping costs (negative), conservation program participation (positive), and income (positive). These results confirmed the stated hypotheses with respect to factor association except for the effect of pumping cost, which was expected to have a positive association with the adoption of conservation practices. It is not clear from the data what drives this result, which is limited to the adoption of surge irrigation only and does not appear statistically significant for other practices.
