Filters
View as a list
Back to map

Implementing a smartphone-based monitoring system in Plant your Future's agroforestry project in the Peruvian Amazon

Submitted by Ralph Ripken 22nd September 2015 1:19
Smartphone based monitoring in agroforestry plot in Peruvian Amazon

Smartphone based monitoring in agroforestry plot in Peruvian Amazon. Measuring DBH and recording information directly in the Open Data Kit Application.

Introduction

Plant your Future aims to create a financially sustainable model for smallholder farmers in the Peruvian Amazon Rainforest through diversifying their income streams resulting in improved livelihoods and increased forest cover. The unique element of our model is access to diversified income streams through products that realize value at different times: fruit, timber and carbon. The project will restore degraded unproductive agricultural lands and reduce deforestation pressure from agriculture and help protect irreplaceable biodiversity. Slash-and burn subsistence agriculture traps farmers in poverty; they face technical, commercial and financial barriers to implementing sustainable agriculture in a more variable local climate.

One to three years after initial planting Plant your Future aimed to implement a smartphone based monitoring system to enable an early record of the agroforestry scheme:

  • To analyse any trends with certain variables such as maintenance, ultimately to improve management (and with further monitoring), to inform future expansion
  • To inform future carbon verification (compare to growth models and calculate an initial biomass estimate)

This case study will, after laying out the climate risks, socio-economic context and project background, describe the monitoring aims, the smartphone based monitoring approach and then discuss practical considerations and limitations.

Climate risks and socio-economic background

Climate Risks

Changes in the local land use scenario due to the impacts of climate change are unlikely – the crops grown for subsistence will still be viable through the project’s lifetime. A map published in Climate change adaptation in Peru: The local experiences (Clements et al., 2010) shows a medium to low confluence of climatic hazards, biodiversity, desert encroachment and poverty – key determinants of the severity of climate change impacts – in Loreto and Ucayali. Although one of the major impacts of climate change in the Amazon region may be flooding, none of the project areas are in floodplains (UNDP, 2013).

On the other hand, 2005 and 2010 experienced large droughts in all of the Amazon region. A 20-­30% increase in fire risk is predicted due to less precipitation and a longer dry season – this could have a negative effect on the ability of farmers in Ucayali and Loreto to grow crops and weather and seasons become less predictables, as well as tend to cattle. Farmers already notice that the rainy season doesn’t start and stop with the regularity that it has in the past. Without the ability to make accurate predictions, they risk not planting crops and trees early enough to get them established before the rainy season ends and the dry season begins.

Pioneer species will change their range to higher altitudes as temperatures at the elevation of the Project Zone rise. Erosion will become a more significant issue with increasingly variable precipitation events. There will be increased human pressure on natural resources. Deforestation will contribute to infrastructure instability.

Socio-economic background of communities

Plant Your Future works with 18 subsistence smallholder farmers in communities who directly directly derive an income from agricultural and forestry produce generated on their land. The communities are located in the Loreto province between Iquitos and Nauta. There are around 42,000 people living between both cities with a 7% annual population growth, distributed in 200 villages (Tello, 2001).

All villages in Plant Your Future´s project area are characterised by population of smallholder farmers who work land which is privately owned and divided into a mosaic of small parcels. Project farmers own an average of 13.5ha each, including a mix of degraded agricultural land and virgin forest (Mateo, 2013). However, only 1-1.5ha are included in the project area per farmer. This area has been deforested and degraded in the past by slash and burn agriculture.

All respondents to the project’s socioeconomic survey were over 40 years of age and the majority (71%) had been educated at least a primary school level. The average family size among beneficiaries was 2­‐3 people.

The smallholder farmer of the project sell their produce exclusively on the Iquitos market and trade is very variable depending on the season. Produce is sold informally which makes it difficult to plan ahead or be confident on pricing. It proved difficult for farmers to estimate their income, but those who could offered a range of 100-270US$ per month - enough to cover their basic needs, but not enough to make major investments. Besides farming, half of the farmers earn wages from small-scale, local activities: fish farms, small stores, raising pigs, caretaking land for others and general labour.

Climate risk impacts on community livelihoods

Poverty is a principle indicator of vulnerability to climate change. The communities in the Project Zone are hugely economically limited as described above. They have a short-­term vision in decision-­making, and are thus wholly unprepared for climate change’s impacts. Policies that incentivize expanding the agriculture border farther into the Amazon region have encouraged growers to cultivate crops that are ill-­suited to the area’s changing climate (Acosta, 2011). When these farms fail, their workers will be forced to subsistence farming.

The impact on physical infrastructure that is likely due to flooding will negatively affect public health efforts, increasing the prevalence of diseases such as malaria and dengue.

The variety of habitats provided in the Project Zone and surrounding regions will be more limited as flooding or droughts cut off corridors between protected areas and there is increasing human pressure on increasingly limited natural resources.

Project background

Plant your Future aims to restore degraded unproductive agricultural lands in the Amazon to forest cover and reduce deforestation pressure from agriculture. Reforestation is a method of managing natural resouces so that they will resist and recuperate from climatic events.

With the help of our project farmers will be able to plant and maintain 1,000 trees each, and restore value to degraded areas of their farms. The project will provide them training to set up nurseries, prepare their land for the reforestation and then maintain the trees once in the ground. In addition Plant your Future will provide support to sell non-timber products at fair prices by engaging with the formal market and supply chains and obtaining certification for products and timber.The project will allow them to sell ‘carbon offsets’ through our VCS carbon certification programme from the carbon stored in the trees as they grow. This will provide three income streams to the farmers at different points in time: from fruits and crops, timber and carbon offset credits.

The money from sale of carbon offsets can be used to help maintain the trees planted through the project, to ensure they can realise the long-term value from orchard fruit tree production and timber.  The project intends to increase Farmers’ income in the short term through the sale of chilli pepper plants and in the long-term through the sale of orchard tree fruits and sustainably harvested timber which they have grown.

The project will reduce deforestation pressure as farmers will be working to restore degraded parts of their farms, rather than clearing new areas of rainforest. This will reduce biodiversity loss and allow for endangered animal and bird species conservation.

Through the agroforestry system Plant your Future aims to increase household incomes and introduce a medium- and long-term vision in natural resource management. The project puts a large emphasis on building farmers capacity to not only technically grow agroforestry systems but also to build their ability to engage in business profitabily. Plant your Future is planning to work with farmers to help them set up cooperatives and engage in formally in business. This should ultimately help to build farmers resiliance to climate change as they have more savings, and are in a better position through increased capacity to deal with changes in weather patterns.

Aims of monitoring system

Plant your Future decided to implement a smartphone based monitoring system to enable an early record of the scheme with two main aims:

  • To analyse any trends with certain variables such as maintenance, ultimately to improve management (and with further monitoring), to inform future expansion
  • To inform future carbon verification (compare to growth models and calculate an initial biomass estimate)

Monitoring with smartphones using the ‘opendata kit collect’ application  provided an efficient method of capturing and uploading data.

Ultimately, we aimed to for a monitoring system which:

  • Produces data in a format and quality suitable for analysis,
  • Is efficient in the field and user friendly, for all those who will recording information (including in future iterations).
  • Technically provides no problems, in using the phones or in uploading information
  • Is easily amendable and can be used by local staff and farmers in Spanish

Monitoring approach

Developing questions

Plant your Future's local staff in Iquitos collected through a workshop with the farmers a list of parameters to measure and record, for the trees (including e.g. tree height, plant health) as well as the plot (inc e.g. soil type, etc). These intitial questions were discussed and improved by Plant your Future staff, the monitoring team and external experts. In addition we developed a first set of classification categories for the answers.

Transposing questions to Open Data Kit Form

Open Data Kit is a web- and application based open-source software that allows to create a web-based survey on the Open Data Kit website that can then be transposed to an application for smartphones. The survey can either be created directly in an Open Data Kit form, or in an Excel based form which allows further option and flexibility. The Excel based form can be uploaded to the Open Data Kit website. The main advantage of the Excel based form is the option to skip questions which makes the monitoring and recording int he field far more efficient.

Uploading form to phones

The ODK Collect application was downloaded to the smartphones and linked to the Plant Your Future ODK server. This allowed to synchronise the forms directly with the smartphones.

Parameters monitored

In order to achieve the aims of the monitoring we recorded a wide range of parameters such as tree height, DBH, maintenance, light conditions, drainage, slope, fertilisation and existance of diseases.

Monitoring in the field

We followed the Standard Operating Procedure (SOP) of Plant your Future and Winrock International for the monitoring of the plots of the 18 smallholder farmers. The monitoring and recording was carried out during three weeks in August 2015 by a team of two smartphone application survey experts, two agroforestry experts and the smallholder farmers. We build the capacity of the local agroforestry experts and some farmers to use to the smartphone based ODK application for future monitoring. For the monitoring we used two waterproof smartphones, a Garmin GPS to locate the center plot and standard agroforestry measuring equipment.

In the first week of the monitoring we tested the survey on the smartphones and improved the survey form daily in order to ensure an efficient and correct monitoring.

Practical considerations, troubleshooting and development of the form

This section will describe considerations and troubleshouting in the development of the Open Data Kit survey form that can be helpful for other projects aiming to implement a smartphone based monitoring system in an agroforestry project.

Consideration 1: Data suitable for analysis:

We added questions where relevant, to ensure we could easily capture the important information in the form.

  1. Choice of precision: Plant your Futue staff and expert advice from Sandra Brown decided which questions required a precise answer and which were suitable to be reported in a category (e.g. DRS < 5cm could be reported as 1-2cm, 2-3cm etc, but DBH and slope should be reported exactly).
  2. Including all relevant information: After a few days testing in the field, it appeared that a number of additional questions would be necessary for analysis. For example:
    • many trees had several stems at 1.30m and thus more than one DBH. We needed space for this in the form;
    • it was necessary to measure the DBH higher than 1.30cm for some trees with wide roots, so we needed to note somewhere how much higher the DBH was measured;
    • some trees were re-planted later than the initial date of planting of the entire plot and we needed to note this, etc.
  3. We added questions where relevant, to ensure we could easily capture the important information in the form. 
  4. Minimise open questions: ODK has the option of ‘select_one’ or ‘select_multiple’ questions. In order to make the data easily usable for analysis, with categories consitent across options, we set as many questions as possible in this way. Where needed (tree species, etc), we included an ‘other’ option. Using categories makes easier analysis by avoiding different spellings of the same option.
  5. Typing errors: In ODK, there is the option to constrain questions, to minimize typing errors. For example, tree height cannot be more than 50m (to avoid a situation where someone notes 98cm as 98m).

Consideration 2: Efficiency in field and user friendly

  1. This choice of categories (rather than having to type text in full, as an open question) also made data collection a little faster in the field.
  2. Skipping questions: As mentioned above, one of the main benefits of using the Excel form over ODK build is the ability to skip questions – to only show questions where relevant. We used this for a number of stages in the form. For example, DRS only shows where DBH is less than 5cm, in line with our SOPs. Space for the DBH of 20 trunks only shows up when the option is selected that the tree has more than 1 trunk. If there is no measurement for a 4th trunk, no more question spaces show up.
  3. Match to information provided in field: Through testing in the field, we tried to match the questions to the information which is shouted out or most easily available.
  4. Images: For the select_one or select_multiple questions, we considered using the ‘image’ column of the form, which would mean that the options would appear as pictures rather than text. However, many of the options were difficult to clearly portray as a small image  (e.g. the small differences between naranja and mandarina were not possible to show as an icon). It was therefore easier for the farmers and those in the field to answer based on the text options.
  5. Ordering trees: Initially, we numbered the trees in the form, to enable a record of the location of each tree (N to S). However, the preceding number was difficult to remember and there were a few errors / instances of having to return and count to check, an inefficient use of time. We also found that the forms did not automatically upload in the order of being written, so there was no way to record this on the excel sheet. We solved this by adding a question to the ODK form: start time (calculated automatically; did not need to be entered), which means we could order the trees later manually in excel.
  6. Recording teams: In order for there to be 2 teams and to recognize who was measuring and recording for each tree (and to avoid this being entered for every form), we added another automatic question: phone ID.

Technical considerations: Using phones in the field and uploading

  1. Working in teams: The phone would be awkward to record if trying to take measurements at the same time. It worked well in a team with 2 or 3 people: one person held the phone and moved through the questions, one person measured the DBH and DRS and called this out, the other person set up for measuring the height.
  2. Phone: Given the wet surroundings we used the waterproof phone SONY M2 Aqua with case added. Humidity and raindrops were no problem. Heavy rain made it however impossible to navigate the touchscreen, unless standing below a tree with big leaves (e.g. banana tree).
  3. Battery: We set the phones to battery conservation mode. Each 30-45mins of usage depleted the battery by about 10%, so by the end of a longer day where the phone was used continuously, the battery needed to be recharged.
  4. Additional applications: We installed OruxMaps on the phones, a GPS map application  that allows using all GPS data for the plots (with significantly more data points than older Garmin GPS units) and locating the center of each sample plot.
  5. Uploading:  There were no problems with uploading forms. With a slow internet connection in the evening, often 2 or 3 attempts would be required, but in the end all non-image forms uploaded to the cloud.

Recommendations

The quality of the form and the monitoring system rests upon:

(i) the quality and choice of the questions and categories, and continually improving to make sure that the right information is captured

(ii) the technical capabilities to make data collection as smooth as possible: this depends upon possible improvements and updates in ODK programmes

(iii) the technical capabilities of the smartphone for monitoring in rainforest settings. There don’t seem to be any major problems with this aspect.

We would recommend for future similar monitoring projects to carefully consider the following aspects:

1. Choice of the questions, categories and wording

  1. Overview & Choice of questions: A thorough reflection is needed on all questions, to decide which questions are necessary for analysis. Consultation with farmers and those in the field, scientific knowlede, plus other projects where data may be possible to cross-compare. It is also important to follow with a process of refinement, to avoid time collecting too many variables in limited field time.
  2. Part of this consideration should be the choice of questions to be answered in the field: e.g. Plot level questions were rarely able to be answered in the field, since they required going back to documents with dates of planting, or soil analysis. For the plot level questions, we would recommend that only the number and farmer name need to be answered in the field - the others would be more effectively simply added in excel afterwards.
  3. Closed questions: A few open questions would be more suitable in the future as closed questions.
  4. Objectivity of categories: Clear categories & descriptions, plus training before the monitoring to make sure everyone has the same, consistent understanding of the categories is essential. This is not only necessary for meaningful analysis, but will also make monitoring easier in the field.

2. Technical capabilities of the smartphone for monitoring in rainforest settings

Battery:  The battery depleted 10% for every 30-45mins usage in the field. This meant there were no problems when away from electricity supply for 3 days in Palo Seco, however this was the maximum time: the batteries were close to 0. With future expansion of the future to rural areas, we would recommend either a battery bank, or uploading the application on additional smart phones with sufficient battery life.

Overall, using the phones with ODK collect seems to be an efficient and appropriate method for monitoring in this setting. This application can be individualised for each monitoring setting, can avoid errors, both in recording and reading/transposing and leads to a fast method of uploading, ready for analysis.

Further Resources