Primary productivity and climate control mushroom yields in Mediterranean pine forests

Highlights

• Previous year primary productivity (NDVI) was the best predictor of mushroom yield.

• Fruiting year precipitacion and temperature codetermined mushroom yield.

• Combining remote sensing and climate data we predicted 55–75% of mushroom yields.

Abstract

Mushrooms play a provisioning ecosystem service as wild food. The abundance of this resource shows high annual and interannual variability, particularly in Mediterranean ecosystems. Climate conditions have been considered the main factor promoting mushroom production variability, but several evidences suggest that forest composition, age and growth play also a role.

Long-term mushroom production datasets are critical to understand the factors behind mushroom productivity. We used 22 and 24 year-long time series of mushroom production in Pinus pinaster and Pinus sylvestris forests in Central Spain to evaluate the effect of climate and forest productivity on mushroom yield. We combined climatic data (precipitation and temperature) and remote sensing data (soil moisture and the Normalized Difference Vegetation Index, NDVI, a surrogate of primary productivity) to model mushroom yields for each forest and for the main edible species of economic interest (Boletus edulis and Lactarius deliciosus).

We hypothesized that mushroom yield would be related to (i) forest primary productivity inferred from NDVI affects mushroom yields, that (ii) soil moisture inferred from remote sensors will equal the predictive power precipitation data, and that (iii) combining climatic and remote sensing will improve mushroom yield models.

We found that (i) previous year NDVI correlated (r = 0.41–0.6) with mushroom yields; (ii) soil moisture from remote sensors rivaled the predictive power of precipitation (r = 0.63–0.72); and (iii) primary production and climate variances were independent, thus the combination of climatic and remote sensing data improved models with mean R²_adj as high as 0.629.

On the light of these results, we propose as a working hypothesis that mushroom production might be modelled as a two step process. Previous year primary productivity would favour resource accumulation at tree level, potentially increasing resources for mycelia growth, climatic conditions during the fruiting season control the ability of mycelia to transform available resources into fruiting bodies.

Introduction

Fungi play key functions in forest ecosystems. Fungi contribute to soil nutrient balance by decomposing organic matter and turning it into inorganic components that are accessible to tree roots. Mycorrhizal fungi also form symbiotic associations that increase trees rhizosphere, eventually improving water and nutrient availability, enhancing tree growth and survival and providing defense against pathogens (Allen, 1991). In addition, mushrooms play a provisioning ecosystem service as wild food that has been acknowledged for a long time across multiple cultures (Boa, 2004). The growing consideration of mushrooms as a delicatessen, with their consequent commercialization, is triggering a transformation on the alimentary sector (Zambonelli and Bonito, 2012). Mushroom supplies are mostly collected in the wild and, as a result, mushroom picking has become a popular leisure activity for urban people. In fact, the development of a mycological touristic sector is having high impact in low-populated, rural areas, contributing to diversify its economy and to expand the touristic season into the mushroom fruiting season (Ágreda et al., 2014; Boa, 2004).

Mycological tourism is compromised by the existence of high uncertainty in wild mushroom yields, which impedes a stable touristic offer (Zambonelli and Bonito, 2012). This phenomenon is particularly acute in environments where climatic conditions show extreme variability among consecutive years, such as Mediterranean ones, since mushroom yields reflect inter-annual climate variations, both in terms of total production and timing of the yield season (Ágreda et al., 2015; Collado et al., 2019). Although forest management can enhance wild mushroom production by promoting tree vigor (Tomao et al., 2017), it does not diminish inter-annual variability driven by weather conditions (Ágreda et al., 2016). As a result, climate change might affect wild mushroom yields, since more intense drought events and higher evapotranspiration may play deleterious effects. However, later mushroom seasons and, particularly, more abundant spring yields due to changes in climate might provide novel windows of opportunity (Büntgen et al., 2012; Sato et al., 2012).

Developing reliable predictive models for mushroom yields is therefore a must for the expansion of this economic sector (Tomao et al., 2017). Indeed, modeling factors that determine wild mushroom yields has become an expanding area of research that benefits from the ever-growing availability of long-term data sets of mushroom yields (Alday et al., 2017; Egli et al., 2006; Fernández-Toirán et al., 2006; Herrero et al., 2019; Martínez-Peña et al., 2012). Weather conditions have been the main environmental factor considered in modeling mushroom yields, temperature being key in temperate forests (Sato et al., 2012) and precipitation in drought-limited Mediterranean environments (Ágreda et al., 2015, 2016; Alday et al., 2017; Herrero et al., 2019). Minimum temperatures can also affect mushroom yields through their effect on fruiting season length (Ágreda et al., 2015). More refined models include forest structure and tree growth rates (Bonet et al., 2008; Herrero et al., 2019), with some attempts to link mushroom yields with series of tree secondary growth (Collado et al., 2019; Primicia et al., 2016). The predictive power of these models is, however, limited and highly dependent on data collected at a local scale.

Remote sensing data have disrupted forest management by being able to monitor forest dynamics at multiple spatio-temporal scales (Barrett et al., 2016), and LiDAR techniques have been proven successful to evaluate mushrooms diversity and production (Moeslund et al., 2019; Peura et al., 2016). Soil moisture content, a critical factor for fungal growth and mushroom production (Karavani et al., 2018), can be inferred from RADAR sensors (Dorigo et al., 2017; Moran et al., 2000; Paloscia et al., 2013) with time series that are available since 1978 (Dorigo et al., 2017). In the same way, remote sensors give information about the Normalized Difference Vegetation Index (NDVI), which is a good estimator of primary productivity (Birky, 2001; Rouse et al., 1973; Wang et al., 2004a) whose interannual variations have been correlated to tree secondary growth at different spatial and temporal scales (Vicente-Serrano et al., 2016). Although preliminary attempts to correlate fungal fruiting phenology and fungal diversity to annual NDVI have been recently undertaken (Andrew et al., 2018, 2019), the relation between NDVI and fungal production has not yet been explored to the best of our knowledge, in spite of the existing well-known positive relationship between forest primary productivity and fungal yields (Ágreda et al., 2014; Alday et al., 2017; Collado et al., 2019; Herrero et al., 2019). Remote sensing data are not independent from climate, since soil moisture responds to precipitation, evapotranspiration and soil characteristics (Entekhabi et al., 1996) and climate is one of the main drivers of primary productivity in terrestrial ecosystems. Therefore, incorporating remote sensors to mushroom yields’ models is a first step towards the future development of detailed predictive models, which will help to boost the mycological touristic sector at different parts of the world. At the same time, this is also an opportunity to explore more in depth the ecological role of environmental drivers on mushroom production, as well as their potential consequences on ecosystem function.

In this study, we benefited from two of the longest time series of fungal production (22 and 24 years), both collected in central Spain. We used climatic (precipitation and temperature) and remote sensing (soil moisture and NDVI) data to model total mushroom yields in wet and dry pine forests, as well as to model the production of the main species of economic interest at each forest type –Boletus edulis Bull. (king bolete) in wet forests and Lactarius deliciosus (L.) Gray (saffron milk cap) in dry forests–. Our main aim was to check whether and which of remote sensing data will allow to predict mushroom yields. Specifically, we hypothesized that (i) forest primary productivity (estimated by NDVI) will have a positive effect on mushroom yields, albeit this effect will vary depending on the trophic guild (saprophytic vs. mycorrhizal), (ii) soil moisture inferred from RADAR sensors will equal the predictive power of traditionally-used precipitation data, and (iii) the combination of climatic and remote sensing data will increase the predictive power of models for mushroom yields.