Plant Invasion and Land Use in the Galápagos Islands

Lab #5

Plant Invasion and Land Use in the Galápagos Islands

Due: November 17, 2008

Lindsay Leonard

Liza Guzmán

For the agricultural zone

1. What relationships exist between farm characteristics (in the attribute table) and the percentguava by farm?

In order to best determine possible relationships between farm characteristics and percent guava present in each farm, we created scatter plots and added a linear trend line. The farm characteristics that were used for comparison with the percent guava are: number of cattle owned by the farm proprietor, percentage of farm used for crop agriculture, percentage of farm in pasture for cattle grazing, and the last year of clearing of the fields on sample farms. Below are the results found for each comparison.

a)Number of cattle owned by the farm proprietor–this characteristic was chosen because cattle are thought to be one of the drivers for the dispersal of guava seeds in Isabela Island. This graph shows a slightly moderate negative relationship between the number of cattle owned by the proprietor and the percent of guava present in the farm. Looking at the graph below, the relationship weakened by an outlier. It appears that the greater the concentration of cattle on a farm, the fewer guava present. This may be due to trampling or grazing by the cattle of both native plants and guava seedlings, not allowing for the growth of guava trees. So although cattle may serve as a key driver in the dispersal of guava, within the farm a greater concentration of cattle means a greater amount of trampling and/or grazing of guava seedlings.

b)Percentage of farm used for crop agriculture–this characteristic was chosen because we expect that farms that are being used for crop farming would be weeded for guava and/or be less suitable for invasion by guava. Farms that are abandoned should than have a greater amount of guava present. This graph shows a clear moderate negative relationship between the percentage of farm used for crop agriculture and the percent of guava present in the farm. Farms that are not being used for crops (this suggests that they are either abandoned or used for purposes other than crops), tend to have a greater percentage of guava than farms that are being widely used for crops (>20%).

c)Percentage of farm in pasture for cattle grazing–this characteristic was chosen because we expect that farms that have a high percent of the area being used for pasture would tend to be cleared of trees and seedlings and may tend to be trampled often by cattle. This graph shows a clear moderate negative relationship between the percentage of farm in pasture for cattle grazing and the percent of guava present in the farm. Therefore, it is expected that there will be less guava present when there is more area within a farm allocated for pasture. The results of this graph are consistent with the results from the correlation between number of cattle and percent guava present in the farms.

d)Last year of clearing of the fields on sample farms - this characteristic was chosen because we expected to find a low presence of guava in farms that have been recently cleared. We theorized that farms that were not recently cleared would be better suited for guava invasion and wouldtherefore have a greater presence of guava. This graphdemonstrates that at time of the study there was a moderate negative relationship between the last year the field was cleared and the percent of guava present in the farm.

1. What, if any, proximity relationships exist between percent guava by field and (a) roads and (b)buildings?

We used the select by location tool to determine those fields that intersected with buildings and/or roads. There appears to be a clear presence of guava along the length of the roads.

However, the opposite appears to be true for concentration of guava with relationship to building location. By looking at our map (1_lguzman.jpg) we found that there is visibly less guava growing near buildings and more guava growing in high concentrations around roads.

Please see attachment for map of relationship between percent guava by field and roads and buildings (1_lguzman.jpg).

1. Are fields or farms next toabandoned fields/farms more likely to contain a higher percentage of guava?

We assigned farms to categories of ‘abandoned’, ‘partially abandoned’ and ‘active’ using the number of cattle owned by the farm proprietor and the percentage of farm used for crop agriculture. All farms that contained no cattle and whose percentage of the farm used for crop agriculture was equal to or less than 10, was assigned to the category ‘abandoned’. Of the remaining farms, those that contained 30 percent or less of crop agriculture and 15 or less cattle were assigned to the category ‘partially abandoned’. All remaining farms were assigned to the category ‘active’. We define any farms with greater than 40% guava as having a high presence of guava.

A total of 3 farms were included in the ‘abandoned category’. All 3 of these farms have a high presence of guava (54%, 62% and 81%), and have owners practicing livelihoods other than agriculture. There are a total of 4 farms in the ‘partially abandoned’ category and all except one of these farms have a high presence of guava (49%, 70% and 75%). Also all 4 farms have owners practicing livelihoods other than agriculture. The one exception is a farm that has only 34% guava. This may be due to this particular farm having 50 percent of its farm in pasture. As seen in the scatterplot for percent guava in pasture, there is a negative relationship between percent of farm in pasture and the amount of guava present. Therefore it is not surprising that although this farm is ‘partially abandoned’ it does not contain a high presence of guava. The remaining 9 farms were placed within the ‘active’ category. Only one of these farms has a high presence of guava (65%). This farm also has 40 cattle with 80% of the farm in pasture and 10% for crops. All the other 8 active farms do not have a high presence of guava and only one of these farms has an owner practicing a livelihood other than agriculture.

A layer was created to show all farms that have a high presence of guava (>40%). A total of 7 farms have a high presence of guava, and all of these except one fall within the categories of ‘abandoned’ and ‘partially abandoned’.

By qualitatively analyzing our map of the relationship between percent guava by field and farm characteristics (2_lguzman.jpg), we found two diagonal swaths with high percent guava (>40%) that fall in a diagonal, southeastern direction. From top to bottom of the leftmost swath, there appears to be an increasing concentration of guava as the field types change from ‘active’ to ‘partially abandoned’ to ‘abandoned’. As mentioned above, this also demonstrates that concentration of guava levels increase towards the southeastern corner of the study area.

Please see attachment for map of relationship between percent guava by field and farm characteristics (2_lguzman.jpg).

Farm_Id / Area / Cattle / Crops / Pasture / Other / Percent Guava / Farm Status
6 / 1332164.63 / 0 / 10 / 0 / Tourism / 53.53 / Abandoned
14 / 948362.59 / 0 / 0 / 0 / Fishing / 81.15 / Abandoned
16 / 762189.74 / 0 / 10 / 0 / Tourism / 61.59 / Abandoned
1 / 1076652.22 / 40 / 10 / 80 / 64.81 / Active
2 / 724830.62 / 30 / 0 / 100 / Tourism / 35.67 / Active
3 / 435215.76 / 20 / 30 / 60 / 19.60 / Active
4 / 1586367.01 / 50 / 20 / 70 / 26.58 / Active
7 / 649346.85 / 20 / 30 / 40 / 38.86 / Active
9 / 325337.84 / 15 / 20 / 60 / 29.79 / Active
12 / 388905.45 / 30 / 30 / 50 / 22.94 / Active
13 / 301101.17 / 30 / 10 / 90 / 12.52 / Active
15 / 204773.43 / 20 / 30 / 50 / 26.28 / Active
5 / 487843.39 / 15 / 0 / 40 / Fishing / 70.34 / Partially Active
8 / 611511.53 / 10 / 10 / 10 / Fishing / 74.41 / Partially Active
10 / 335242.36 / 0 / 30 / 0 / Tourism / 49.24 / Partially Active
11 / 205455.94 / 10 / 10 / 50 / Fishing / 34.19 / Partially Active

1. Is there spatial clustering of guava in the agricultural zone or in individual fields, and does thisspatial clustering provide clues to land use practices that could affect guava distribution?

From visually looking at the map, ‘Guava clusters by field in farm’ (3_lguzman.jpg) it is clear that there are 4 spatial clusters of guava in the agricultural zone. One cluster in the northwest falls within farm_id 1;there is a cluster in the NE that falls within farm_id 8; a third cluster is visible in the south east side and falls within two farms: farm_id 14 and 16; there is also a small cluster in farm-id 17. By comparing the previous map, 2_guzman.jpg, we found that these clusters fall within farms that are not actively being used for agricultural purposes or are abandoned. We also compared 3_lguzman.jpg to 1_lguzman.jpg and found that the southeastern cluster also falls in an area where farm fields are adjacent to roads. These land use practices appear to be a contributing factor in the ability for guava to invade a given area.

Please see attachment for map of guava clusters by field in farm (3_lguzman.jpg).

For the Park zone

1. Is there a relationship between guava in the Park and distance to the agricultural zone?

Yes, there is a relationship between guava in the park and distance to the agricultural zone. Using spatial analysis, we found that as distance to the farm boundary increasesthe amount of guava decreases. Guava appears to be an invasive species that is growing past the boundary of the agricultural zone—where it is centered—and is spreading into the park-only zone. We conclude that because of this there is a relationship such that guava concentration decreases as distance from the source increases. Please note that classes (1-10) increases with distance from agriculture zone.

Please see attachment for map showing distance from park to field (4_lguzman.jpg).

1. Is there a relationship between guava distribution in the Park and percent guava in fieldsbordering the park?

Yes, there is a relationship between guava distribution in the park and percent guava in fields bordering the park. In order to determine if there is a relationship between the two we used attribute data from farm_fields.shp. For our map (5_lguzman.jpg), we found that fields within the 500m band that have the highest concentration of guava are located in the southeastern quadrant. These areas are adjacent to fields that are classified as (1) having high concentrations of guava (>40%) and (2)are ‘abandoned’.

The one exception to the trend where guava increases in a southeastern direction lies in the most northeastern area of the study site (FID 48). This field was last cleared in 1990, which supports part (d) of question 1 where we saw a clear negative relationship between percent guava and the last year the field was cleared. Also this field falls within a farm classified as ‘partially abandoned’. The percent guava for this farm is 74%, while the percent guava within the field is equal to 84.93%. This means that the presence of guava in this one farm is clustered within the field (FID 48) which lies adjacent to the boundary of the park-zone with the highest concentration of guava within the 500m boundary (#50).

Please see attachment for map showing relationship between 500m band and farm_field attributes (5_lguzman.jpg).

1. How about between guava distribution in the park and characteristics of farms bordering the Park?

Yes, there is a relationship between guava distribution in the park and characteristics of farms bordering the park. Using data from farms.shp, we found that with one exception, guava concentration within the band begins to decrease in a northwestern direction the further it gets from abandoned or partially abandoned fields. The one exception to this trend are the three farms classified as ‘active’ that lie immediately north of the most southeastern portion of the 500m band. These border fields within the 500m band (11,12,14) that have a high concentration of guava. We hypothesize that this exception is due to (1) a trend where concentration of guava increases in a southeastern direction and (2) that the area of these farms is small and does not provide a very wide buffer between two areas with high guava concentration.

The southwestern farms that are classified as ‘active’ are another example of areas that could have the potential for high guava concentration if they were not currently being used for agriculture. These farms cover a larger area than the southeastern ‘active’ farms mentioned above and because of thisthey may act as a buffer between areas of high guava concentration.

Please see attachment for relationship between 500m band and farm attributes (6_lguzman.jpg).

Story Summary

Using quantitative and qualitative analysis as well as other statistical methods we were able to reach a more scientific understanding of guava invasion on Isabela Island in the Galapagos, Ecuador. Most importantly we found relationships between percent guava and farm characteristics, field characteristics, and distance from the agricultural zone to park-only areas.

For the agricultural zone we looked at farm and field attributes. Farm characteristics included: (a) number of cattle owned by the farm proprietor, (b) percentage of farm used for crop agriculture, (c) percentage of farm in pasture for cattle grazing, and (d) the last year of clearing of the fields on farms. While cattle are believed to be one driver in guava seed dispersal, we found there to be a tipping point where too many cattle decreased the concentration of guava—which is possibly due to excess trampling and/or grazing of guava seedlings. Also, there is less guava present when there is more area in a farm allocated for pasture and cattle grazing. In addition we found that farms not being used for crops tend to have a greater percentage of guava than farms that are actively being used for agriculture. We also found a negative relationship between the last year of clearing and percent of guava per area.

We analyzed field attributes and their relationship to patterns in concentration of guava within each farm. There is visibly less guava growing near buildings and more guava growing in high concentrations around roads. There is also a high presence of guava in farms that we classified as ‘partially abandoned’ or ‘abandoned’ and our data supports our theory that guava increases as field types change from ‘active’ to ‘partially abandoned’ to ‘abandoned’, respectively.

In order to study the relationship between the park-only and agricultural zones, we looked at patterns in concentration of guava within the farms/fields and their distance from the park boundary. Guava levels within the park-only zone decrease as distance from the agricultural zone increases. Areas within the park that have the highest guava concentration fall adjacent to farms/fields that are classified as having high guava concentration and ‘abandoned’ or ‘partially abandoned’. When creating fields within a 500m band outside of the agricultural zone, we discovered a clear trend where guava concentrations increase in a diagonal, southeastern direction.

We would be interested in seeing data south of the given data to further support our hypothesis of a southeastern trend. It might also be good to have additional data that describes grazing habits of cattle not only within the agricultural zone but also within the park-only zone in order to see their effect on guava concentration. Finally, we would also like to look at wind patterns as a driver in seed dispersal.

Appendix 1

Table of farm_field characteristics with percent guava

ID / VALUE_1 / % guava / Farm_Id / Area / Cleared / Pasture / Crops
1 / 4377.60 / 4.14 / 4 / 105834.61 / 2004 / 1 / 0
2 / 45141.12 / 43.08 / 4 / 104791.61 / 2000 / 0 / 1
3 / 22314.24 / 25.82 / 4 / 86418.01 / 2003 / 0 / 1
4 / 29669.76 / 18.71 / 4 / 158542.82 / 2003 / 1 / 0
5 / 8876.16 / 10.23 / 4 / 86741.07 / 2001 / 1 / 0
6 / 38534.40 / 17.90 / 4 / 215252.90 / 2000 / 1 / 0
7 / 23408.64 / 17.57 / 4 / 133249.52 / 2001 / 1 / 1
8 / 11888.64 / 7.55 / 4 / 157532.75 / 2001 / 1 / 0
9 / 4072.32 / 8.46 / 8 / 48146.23 / 2003 / 1 / 1
10 / 5408.64 / 16.60 / 8 / 32577.40 / 2003 / 0 / 1
11 / 19272.96 / 30.89 / 10 / 62395.40 / 2004 / 0 / 1
12 / 23351.04 / 32.82 / 11 / 71148.34 / 1999 / 1 / 0
13 / 59345.28 / 74.86 / 10 / 79279.91 / 1998 / 0 / 1
14 / 19100.16 / 28.86 / 10 / 66170.98 / 1999 / 0 / 1
15 / 67351.68 / 52.87 / 10 / 127396.07 / 1995 / 0 / 0
16 / 10644.48 / 12.77 / 11 / 83371.31 / 2000 / 1 / 1
17 / 36259.20 / 71.19 / 11 / 50936.30 / 1999 / 1 / 0
18 / 64298.88 / 88.41 / 9 / 72729.45 / 1990 / 0 / 0
19 / 4999.68 / 8.92 / 9 / 56043.50 / 2004 / 0 / 1
20 / 13979.52 / 20.35 / 9 / 68683.89 / 2001 / 1 / 0
21 / 8893.44 / 9.80 / 9 / 90722.59 / 2002 / 1 / 0
22 / 684495.36 / 70.82 / 1 / 966510.84 / 1992 / 1 / 0
23 / 13484.16 / 12.24 / 1 / 110141.37 / 2003 / 1 / 1
24 / 311253.12 / 86.27 / 5 / 360801.82 / 1990 / 1 / 0
25 / 157495.68 / 61.75 / 4 / 255052.59 / 2002 / 1 / 0
26 / 79701.12 / 28.15 / 4 / 283132.19 / 2001 / 1 / 1
27 / 31893.12 / 25.10 / 5 / 127041.57 / 1999 / 1 / 0
28 / 237363.84 / 60.83 / 6 / 390189.33 / 1995 / 0 / 0
29 / 258566.40 / 35.67 / 2 / 724830.62 / 1999 / 1 / 0
30 / 85305.60 / 19.60 / 3 / 435215.76 / 2001 / 1 / 1
32 / 4740.48 / 12.74 / 9 / 37211.25 / 2002 / 1 / 0
33 / 10373.76 / 10.96 / 12 / 94688.80 / 2002 / 1 / 0
34 / 44179.20 / 37.20 / 12 / 118747.60 / 2000 / 0 / 1
35 / 3974.40 / 8.51 / 12 / 46728.43 / 2003 / 1 / 0
36 / 2188.80 / 3.71 / 12 / 58948.26 / 2004 / 1 / 0
37 / 14117.76 / 80.84 / 12 / 17464.52 / 1995 / 0 / 0
38 / 3133.44 / 41.75 / 12 / 7504.96 / 1995 / 0 / 0
39 / 9365.76 / 55.78 / 12 / 16791.90 / 2000 / 0 / 1
40 / 1900.80 / 6.73 / 12 / 28259.84 / 2001 / 1 / 0
41 / 37699.20 / 12.52 / 13 / 301101.17 / 2000 / 1 / 1
42 / 396201.60 / 76.53 / 16 / 517723.74 / 1990 / 0 / 1
43 / 18725.76 / 36.34 / 16 / 51529.35 / 1995 / 0 / 0
44 / 25079.04 / 20.02 / 16 / 125288.51 / 1995 / 0 / 0
45 / 29404.80 / 43.47 / 16 / 67648.14 / 1995 / 0 / 0
46 / 53809.92 / 26.28 / 15 / 204773.43 / 2000 / 1 / 1
47 / 252322.56 / 38.86 / 7 / 649346.85 / 1999 / 1 / 1
48 / 475776.00 / 50.51 / 6 / 941975.30 / 1990 / 0 / 1
49 / 769605.12 / 81.15 / 14 / 948362.59 / 1985 / 0 / 0
50 / 445530.24 / 83.93 / 8 / 530809.39 / 1990 / 0 / 0